{"id":"716360ac-b554-4748-b796-2fd91b82c5af","arxiv_id":"1909.00809","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":3.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A perspective review that attributes defectiveness in non-Hermitian lattices to boundary conditions of a hypothetical Hermitian parent system.","lead":"This paper reviews the emerging field of topological states in non-Hermitian lattices, where energy can be complex. It proposes a scattering-picture interpretation for the strange lack of some eigenstates, called defectiveness.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Scattering-parent interpretation is only demonstrated for real-energy resonance eigenstates, so its scope over generic defective non-Hermitian lattices is unsupported.","rationale":"The reader's verdict CONDITIONAL is appropriate. The paper is a Perspective whose review portions are accurate, and the author explicitly acknowledges the parent-existence caveat; my concern is not an internal inconsistency but a scope gap. I agree with the reader's weakest_assumption: the load-bearing premise is the existence and completeness of a Hermitian parent. I would not change the verdict, because the paper's framing is appropriately conditional and useful as an entry point. The concrete test could sharpen the condition by determining empirically whether the mapping covers skin-effect models, the flagship phenomenon in Section 6.2. The paper deserves credit for flagging its own main limitation, and the interpretive claim is presented as a perspective rather than a theorem.","tokens_in":21256,"tokens_out":5299,"duration_ms":61316,"concrete_test":"Apply the Jin–Song embedding of Section 6.1 to the one-dimensional Hatano–Nelson chain with asymmetric hoppings t_L ≠ t_R and open boundaries: construct the natural Hermitian parent (the same lattice with the imaginary on-site term replaced by a semi-infinite lead), solve the scattering problem with both left- and right-incident boundary conditions, and compute the projected effective Hamiltonian on the common Hilbert space. Then compare its eigenvalues and eigenvectors with the full open-boundary spectrum of the non-Hermitian chain. If the parent reproduces only real-energy resonances, or only a subset of states, while the complex-energy skin modes have no parent counterpart, the claimed d→d+1 correspondence fails for the very models used to motivate the skin-effect discussion.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central interpretive claim—defectiveness as lack of boundary-compatible states, and the d→d+1 Hermitian-parent correspondence—requires that a non-Hermitian lattice be obtainable, as a whole, by projecting a Hermitian parent plus a boundary condition. The paper's own Section 6 explicitly concedes that 'there is no warranty that a non-hermitian Hamiltonian can always be assimilated to the effective description of a scattering situation,' and the footnote adds that even when a parent exists the effective description is energy dependent, so the link is restricted to discrete resonance energies. Section 6.1 then shows only a subset correspondence: real-energy eigenstates of PT-symmetric tight-binding models map to resonant transmission states of a Hermitian parent, and the time-reversed partner belongs to a different child (opposite sign of the imaginary potential). Thus each non-Hermitian child is defective because its TRS partner was assigned to the other child, not because a generic defective Hamiltonian has a parent. No construction or numerical evidence is given for models with genuinely complex spectra, such as the Hatano-Nelson asymmetric-hopping chains that exhibit the skin effect discussed in Section 6.2. The dimensionality link is therefore an expectation with a narrow demonstrated base; it has not been shown to preserve the full spectrum, eigenvectors, or topological invariants. Because the author flags this limitation, the paper is honest, but the central claim's scope is exactly as wide as the unproven parent-existence condition.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This Perspective surveys recent work on topological states of non-Hermitian lattices. It reviews possible definitions of a gap in the complex energy plane (point gap and line gap), the meaning of imaginary eigenvalues, the physical origin of gain, loss, and asymmetric couplings, and the role of exceptional points and defectiveness. The paper's central proposal is that a non-Hermitian lattice can be understood as the projection of a Hermitian parent lattice in one higher dimension, with a boundary condition that selects a subset of scattering states; defectiveness is then interpreted as the absence of enough eigenstates compatible with that boundary condition. This idea is illustrated with the Jin-Song construction for PT-symmetric tight-binding chains and with a Chern-ribbon scattering picture, and is connected to the non-Hermitian skin effect. The paper also surveys proposals for non-Hermitian bulk-boundary correspondence and analogies with Floquet systems. It explicitly acknowledges that not every non-Hermitian Hamiltonian admits such a Hermitian parent.","tokens_in":21419,"tokens_out":7013,"duration_ms":72434,"significance":"If the scattering-parent interpretation holds for the intended class of systems, it offers a unifying physical picture of defectiveness and skin effects, and it connects non-Hermitian topology to familiar Hermitian scattering and Floquet concepts. The paper is a useful perspective: it gives a clear account of point-gap versus line-gap notions, summarizes competing routes to bulk-boundary correspondence, and is unusually candid about the main weakness of its own proposal. It does not claim to fit parameters or derive new invariants, and it makes no falsifiable predictions; its contribution is conceptual synthesis. The explicit self-criticism in Section 6 is a strength, because it prevents the interpretive claim from being mistaken for a theorem.","major_comments":[{"comment":"The central interpretive claim that defectiveness 'can be understood as the lack of enough states which are compatible with the boundary condition' (Section 6) is demonstrated only for a restricted class: real-energy eigenstates of PT-symmetric tight-binding models that map to resonant transmission states of a Hermitian parent, as shown in Section 6.1. The footnote to Section 6 concedes that even when a parent exists the effective description is energy dependent and the link is restricted to resonances, and Section 6.2 discusses skin-effect models with complex spectra for which no parent construction is given. Please either restrict the claim to the established class or explicitly promote the general statement to a conjecture and state what evidence would confirm or refute it, for instance a parent construction for an asymmetric-hopping chain with genuinely complex eigenvalues.","section":"Section 6 and Section 6.1"},{"comment":"The expectation that 'a non-Hermitian lattice in dimension d is an effective description of an Hermitian lattice in dimension d + 1 plus boundary conditions' is presented as a consequence of the scattering picture, but the supporting references do not establish the existence of a parent for generic models or the preservation of the full spectrum, eigenvectors, and topological invariants under projection. The paper should either mark this dimensionality correspondence as an open conjecture, with a precise statement of the class of non-Hermitian Hamiltonians to which it is intended to apply, or provide a concrete worked example, such as deriving a non-Hermitian SSH chain with non-reciprocal hoppings from a Chern-insulator ribbon parent plus a boundary condition.","section":"Section 6, dimensionality link"},{"comment":"The argument that each non-Hermitian child is defective at the resonant energy because its time-reversed partner is assigned to the other child is specific to the two-child construction with opposite imaginary potentials. As written, the surrounding text suggests a general mechanism for defectiveness in arbitrary non-Hermitian Hamiltonians. Please label this mechanism as illustrative rather than universal, and clarify that generic defective Hamiltonians may become defective through other mechanisms, such as the higher-order exceptional points discussed in Section 6.2.","section":"Section 6.1, two-child construction"}],"minor_comments":[{"comment":"The dimensionality convention is used inconsistently: the text first maps a d-dimensional lattice to a non-Hermitian lattice in dimension d-1, then states the expectation as a d-dimensional non-Hermitian lattice arising from a d+1-dimensional Hermitian parent. Please fix the notation so the reader can follow the dimensional shift.","section":"Section 6, first paragraph"},{"comment":"The phrase 'This anomalous localization was a attributed to the proximity' contains a typo; it should read 'was attributed to the proximity'.","section":"Section 6.2"},{"comment":"The phrase 'representing containing an absorbing on-site term' should be reworded, for example 'representing a tight-binding network containing an absorbing on-site term'.","section":"Figure 4 caption"},{"comment":"The phrase 'an integral over the the Brillouin zone' contains a duplicated article; it should read 'over the Brillouin zone'.","section":"Section 7"},{"comment":"The crucial caveat about the absence of a guarantee that a Hermitian parent exists, and about the energy dependence of the effective description, is important enough to be stated in the main text; consider moving it out of the footnote.","section":"Footnote in Section 6"}],"recommendation":"major_revision","confidential_remarks":"This is a Perspective rather than a full research paper, so the bar should be whether the conceptual synthesis is useful and honestly labeled. The manuscript is candid about its main limitation, and I found no circularity or fitted parameters. The main risk is that the scattering-parent interpretation is stated more broadly than the demonstrated examples support; a strengthened scope statement should resolve this. The editors may also wish to ask the author to add a brief note on developments since the original submission, given the rapid evolution of the non-Hermitian topology field."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing to know: this is a Perspective, not a primary research paper. The scattering-picture interpretation of defectiveness—defectiveness as missing boundary-compatible states—is explicitly credited to Jin and Song [96,97], and the d→d+1 Hermitian-parent link is presented as 'reasonable to expect' without proof. Do not read it as a solution to the non-Hermitian bulk-boundary correspondence problem.\n\nWhat it does well: it is a clean map of the field's current vocabulary. The point-gap/line-gap distinction is useful, the skin effect and the competing routes to a bulk-boundary correspondence (non-Bloch invariants, Green functions, doubled Hamiltonians) are summarized accurately and cited fairly. The author flags the central weakness himself in Section 6: no warranty exists that a non-Hermitian Hamiltonian admits a Hermitian parent, and even when one exists the effective description is energy-dependent, so the link holds only at discrete resonance energies. That honesty is real credit.\n\nWhere it goes soft: the demonstrated base is narrow. Following Jin and Song, Section 6.1 shows that real-energy eigenstates of PT-symmetric tight-binding models correspond to resonant transmission states of a Hermitian parent, with the time-reversed partner belonging to the other child. That explains why each child is defective, but it does not cover generic defective lattices with complex spectra, such as the Hatano-Nelson chain discussed later. The d→d+1 dimensionality claim is an expectation, not a derivation. As long as the paper says 'when such parent lattice exists,' the claim is safe, but the scope is exactly as wide as that condition—which is unproven generically. The stress-test note is right about this.\n\nThe paper also does not clearly delineate what is new relative to Jin-Song, but since it credits them, I'd treat that as a review-style choice rather than an omission. No data or code are involved, which is expected for a Perspective.\n\nWho this is for: a newcomer or non-specialist wanting orientation in non-Hermitian topology. A practitioner will not find new math or a new prediction. It deserves peer review as a Perspective, not as a research letter. If I were editor, I'd send it out: the author is a serious person, the review is accurate, and the honest framing of the conjecture is exactly what a Perspective should do.","headline":"A candid, well-organized Perspective whose main interpretive suggestion is honestly labeled as a conjecture; worth a referee for a review venue, but don't expect a new result.","tokens_in":22021,"tokens_out":3521,"would_cite":true,"duration_ms":34865,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper argues that a non-Hermitian lattice in dimension $d$ can be read as the projection of a Hermitian lattice in dimension $d+1$ plus a boundary condition, making defectiveness a boundary-condition phenomenon.","keywords":["non-Hermitian lattices","topological states","defectiveness","exceptional points","non-Hermitian skin effect","bulk-boundary correspondence","scattering picture","Floquet systems"],"falsifier":"Pick a canonical non-Hermitian lattice with skin modes, such as the one-dimensional chain with asymmetric hoppings, and try to realize it as the projection of scattering states of a two-dimensional topological ribbon with a chosen incidence direction. If no parent geometry and boundary condition reproduces the skin-mode eigenstates, or if the correspondence holds only at isolated energies that do not cover the whole spectrum, the central claim is falsified. The calculation is concrete: solve the parent scattering problem, project onto the strip, and compare the projected subspace with the child's defective eigenspace.","tokens_in":20980,"feed_emoji":"🔬","tokens_out":12698,"duration_ms":114437,"temperature":0.7,"pith_summary":"This Perspective proposes an interpretive key for topological states of non-Hermitian lattices: a non-Hermitian lattice in dimension $d$ can be understood as an effective description of a Hermitian lattice in dimension $d+1$ plus a boundary condition, obtained by projecting the parent's scattering states onto a sub-Hilbert space. On this view the two features that make non-Hermitian topological physics strange—defectiveness (missing eigenvectors) and extreme sensitivity to boundary conditions—are two sides of the same coin: defectiveness is the lack of parent eigenstates compatible with the chosen boundary condition, and the much-studied skin effect is its extreme case. The author draws on a scattering picture, in which gains and losses in the effective Hamiltonian encode where particles enter and leave, and connects the framework to exceptional points, bulk-boundary correspondence, and Floquet systems. A sympathetic reader would care because if the picture is right, many non-Hermitian phenomena stop being isolated curiosities and become boundary-condition constraints inherited from a hidden Hermitian ancestor.","feed_headline":"Non-Hermitian lattices may be shadows of a hidden Hermitian parent","feed_subtitle":"Defectiveness becomes a boundary-condition effect, linking skin modes, exceptional points, and bulk-boundary failure.","key_machinery":"The load-bearing mechanism is the Hermitian-parent construction: view a non-Hermitian Hamiltonian as the projection, onto a sub-Hilbert space, of a Hermitian parent lattice in one higher dimension together with a boundary condition such as an incidence direction for incoming particles. The boundary condition generates the gains and losses of the effective non-Hermitian child, and defectiveness is the exclusion of parent eigenstates that do not fit that boundary condition. A companion device is the doubled Hamiltonian built from $H$ and $H^\\dagger$ as off-diagonal blocks, which restores the information missing from the non-Hermitian child; in the scattering picture the skin effect appears as extreme defectiveness at exceptional points whose order scales with system size.","core_discovery":"The paper's central claim is that defectiveness—the lack of a full set of linearly independent eigenvectors, the defining non-Hermitian feature at exceptional points—has a concrete physical meaning in a scattering picture. Take a Hermitian lattice with at least one translationally invariant direction, impose a boundary condition such as particles incident from one side, and project the resulting scattering states onto a sub-Hilbert space: the effective Hamiltonian for that subspace is non-Hermitian, with gains and losses that encode the boundary condition. The missing eigenvectors are precisely the parent eigenstates that are incompatible with that boundary condition; they are not absent from the parent, they are excluded by the constraint. The author illustrates this with a ribbon of a two-dimensional topological insulator whose boundary condition selects an edge state at a single edge, and cites the result that real-energy eigenstates of parity-time-symmetric non-Hermitian tight-binding chains correspond to resonant transmission states of a Hermitian parent. The paper also draws the corollary that a non-Hermitian lattice in dimension $d$ is, when a parent exists, the 'shadow' of a Hermitian lattice in dimension $d+1$ plus boundary conditions, and that the link may hold only at discrete resonance energies.","pith_inferences":["If the shadow picture holds generally, it suggests a constructive search strategy for missing parent lattices: for any anomalous non-Hermitian model, look for a Hermitian lattice in one higher dimension whose scattering subspace, after projection, reproduces the model; this would turn classification questions into a search over parent geometries.","The paper's own caveat that the parent link often holds only at resonance energies implies the picture is most secure for effectively single-particle, non-interacting settings; in interacting or time-dependent many-body systems the parent may need to be nonlocal in energy or time, which is a testable limitation rather than a contradiction.","A direct experimental test could be built in photonic or acoustic lattices: tune a boundary condition, such as the side from which a waveguide mode enters, and watch whether the number of linearly independent localized modes in the projected system changes exactly as the scattering picture predicts; appearance or disappearance of defectiveness under boundary-condition tuning would confirm the mech"],"forward_implications":["If the scattering picture is right, the non-Hermitian skin effect is a boundary-condition phenomenon: open-boundary eigenstates localize at one edge because the chosen boundary condition excludes the extended states of the parent, so bulk and finite-system spectra can differ even in the thermodynamic limit.","Complex spectra force a choice of gap definition, and the choice changes the topology: point gaps allow a single band to wind around a base energy and carry a nonzero winding number, whereas line gaps preserve a more Hermitian-like classification.","Bulk-boundary correspondence cannot be assumed; the paper surveys non-Bloch invariants, real-space invariants, Green's-function invariants, and the doubled-Hamiltonian reduction as competing routes to restore it.","Floquet systems become a natural testing ground: projecting Floquet space onto a single replica produces effective non-Hermitian-like behavior, and tools such as the doubled Hamiltonian transfer between driven and non-Hermitian problems.","Defectiveness at exceptional points can be extreme in pristine lattices; the picture frames higher-order exceptional points and eigenspace condensation as the boundary-condition constraint taken to its limit."],"supporting_citations":[{"why":"It supplies the explicit construction: real-energy eigenstates of a non-Hermitian tight-binding chain match resonant transmission states of a Hermitian parent with a lead.","marker":"[96]"},{"why":"It extends the parent construction beyond the parity-time-symmetric case and gives the projection recipe used to interpret defectiveness.","marker":"[97]"},{"why":"It shows that perfect-transmission scattering can be formulated as a parity-time-symmetric spectral problem, supporting the scattering picture.","marker":"[98]"},{"why":"It grounds the absorbing-site starting point by treating an on-site imaginary potential as an effective description of coupling to an outer world.","marker":"[102]"},{"why":"It reports anomalous localization and eigenspace condensation at exceptional points, which the scattering picture interprets as boundary-condition incompatibility.","marker":"[46]"},{"why":"It introduces and analyzes the non-Hermitian skin effect, the extreme defectiveness that the proposed picture aims to explain.","marker":"[47]"},{"why":"It characterizes skin modes and their existence conditions, a central target phenomenon for the interpretation.","marker":"[64]"},{"why":"It shows that a one-dimensional chain with asymmetric hoppings is the non-Hermitian proxy of a topological insulator, supporting the claimed dimensionality link.","marker":"[50]"},{"why":"It documents the failure of bulk-boundary correspondence and the extreme sensitivity of the spectrum to boundary conditions, motivating the scattering interpretation.","marker":"[45]"}],"fun_headline_variants":["Non-Hermitian lattices: shadows of a hidden Hermitian parent","Defectiveness as a boundary-condition effect in non-Hermitian lattices","Non-Hermitian lattices are boundary-condition shadows","Boundary conditions expose hidden Hermitian parents"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a non-Hermitian Hamiltonian can always be viewed as the projection of a Hermitian parent Hamiltonian plus a boundary condition; the paper explicitly concedes that no such parent is guaranteed to exist and that, when it does, the link often holds only at discrete resonance energies.","fun_headline_variants_meta":{"raw":{"variants":["Non-Hermitian lattices: shadows of a hidden Hermitian parent","Defectiveness as a boundary-condition effect in non-Hermitian lattices","Non-Hermitian lattices are boundary-condition shadows","Boundary conditions expose hidden Hermitian parents"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001569,"raw_usage":{"total_tokens":6225,"prompt_tokens":865,"completion_tokens":5360,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":481,"completion_tokens_details":{"reasoning_tokens":5290}},"tokens_in":481,"tokens_out":5360,"duration_ms":37410,"temperature":1.0,"reasoning_tokens":5290,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T05:34:53.961825+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Pick a canonical non-Hermitian lattice with skin modes, such as the one-dimensional chain with asymmetric hoppings, and try to realize it as the projection of scattering states of a two-dimensional topological ribbon with a chosen incidence direction. If no parent geometry and boundary condition reproduces the skin-mode eigenstates, or if the correspondence holds only at isolated energies that do not cover the whole spectrum, the central claim is falsified. The calculation is concrete: solve the parent scattering problem, project onto the strip, and compare the projected subspace with the child's defective eigenspace.","supporting_citations":[],"review_version":1}