{"id":"da94c32e-08ea-4853-bdd9-09558dad1eb9","arxiv_id":"1908.08434","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Discrete spectrum, bounded measure complexity, and the two mean-equicontinuity notions coincide for invariant measures of countable amenable group actions, with the equicontinuity parts requiring tempered Følner sequences.","lead":"This paper proves that for actions of countable 'amenable' groups (a broad family including all abelian, finite, and solvable groups), an invariant measure has discrete spectrum exactly when its measure complexity is bounded, and that this is equivalent to two mean-equicontinuity conditions. It transfers a package of equivalences already known for integer-time systems to every countable amenable group.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Arithmetic slip in Lemma 2.5 blocks the written proof of a key implication, but the intended correction is straightforward.","rationale":"Most of the paper's architecture is sound: the reduction of arbitrary Følner sequences to tempered subsequences is legitimate, Lemma 2.2's use of Lindenstrauss's pointwise ergodic theorem is valid for the semimetrics involved, and the pigeonhole/Følner argument in Theorem 2.7 is internally consistent. The one genuinely load-bearing written defect I find is the arithmetic slip in Lemma 2.5. This lemma is the step that converts non-almost-periodicity of h into non-almost-periodicity of the difference function L(x,y)=h(x)-h(y); without it Theorem 2.7 cannot run. The printed estimate after applying Lemma 2.4 has a constant term 4C that does not vanish as k grows, so the final 'by arbitrariness of k' inference fails. The intended bound is (4C+4C^2)/k + 4C^2/(k-1), which does tend to 0 and makes the argument work. Because this is a transparent arithmetic correction rather than a structural flaw, the central claim is very likely true and the verdict remains conditional. The reader's weakest-assumption analysis pointed at Lindenstrauss's theorem and the tempered hypothesis; that is a legitimate boundary of the method, but it is not where I find the concrete error. Hence partial agreement with the reader's assessment.","tokens_in":16665,"tokens_out":34772,"duration_ms":333498,"concrete_test":"Recompute the last two inequalities of Lemma 2.5 by substituting the Lemma 2.4 bound with C replaced by 2C. If the resulting constant is (4C+4C^2)/k + 4C^2/(k-1), then for every ε>0 one can choose k large enough so that the L2 distance ‖h∘g_{k,m}-h∘g‖2 is below ε, and the lemma is valid. If the printed constant 4C + 4C^2/k + 4C^2/(k-1) is retained, exhibit C=1/4 and show the bound cannot be driven below any positive threshold, so the proof of Lemma 2.5 fails as written.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 2.5 is needed for Theorem 2.7 (bounded complexity implies almost periodicity): it shows that if h is not almost periodic, then L(x,y)=h(x)-h(y) is not almost periodic. The displayed estimate before the final sentence reads, with |h|≤C and ∫(h∘g_{k,m}-h∘g)=0, that ∫|h∘g_{k,m}-h∘g|^2 dµ ≤ 2C∫|h∘g_{k,m}-h∘g| dµ ≤ 4C + 4C^2/k + 4C^2/(k-1). Taken literally, the right-hand side tends to 4C, not to 0, so the claim 'by arbitrariness of k>1 we have that h is almost periodic' does not follow. The preceding application of Lemma 2.4 gives ∫|h∘g_{k,m}-h∘g| dµ ≤ (2+2C)/k + 2C/(k-1); multiplying by 2C yields (4C+4C^2)/k + 4C^2/(k-1), which does tend to 0. Thus the proof is salvageable by an evident correction, but as written this lemma has a genuine gap in the converse direction of the main equivalence.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies invariant probability measures for actions of countable discrete amenable groups on compact metric spaces. It defines measure-theoretic complexity functions with respect to Følner averages of a semimetric and with respect to Hamming distances of partition names, and proves Theorem 1.1: for any invariant measure and any Følner sequence, discrete spectrum is equivalent to bounded complexity with respect to the given metric, to bounded complexity for every finite partition, and to bounded complexity for every two-element partition; when the Følner sequence is tempered, these are also equivalent to mean equicontinuity and equicontinuity in the mean. The technical core, Theorem 2.1, characterizes almost periodic L² functions by bounded complexity with respect to the semimetric H(x,y)=|h(x)-h(y)|. The proofs are written out in full, with explicit constants and with careful use of Lindenstrauss's pointwise ergodic theorem and Zimmer's characterization of discrete spectrum via almost periodic functions.","tokens_in":16689,"tokens_out":21440,"duration_ms":182836,"significance":"If correct, the paper gives a genuine extension to countable amenable group actions of the known equivalences for Z-actions among discrete spectrum, bounded measure complexity, bounded partition complexity, mean equicontinuity, and equicontinuity in the mean. This unifies several existing results (Huang--Wang--Ye, Huang--Li--Thouvenot--Xu--Ye, Ferenczi, Vershik--Zatitskiy--Petrov) in a natural generality. The proof strategy is sound: Section 2 develops a self-contained almost-periodicity criterion via complexity, and Section 3 assembles the full equivalence through semimetric and partition arguments. The paper is also careful about the role of tempered Følner sequences, reserving the mean-equicontinuity statements for the tempered case. The main defect is a single local arithmetic slip in Lemma 2.5, which is load-bearing for the converse direction but is straightforwardly corrected.","major_comments":[{"comment":"The displayed estimate before the final sentence of Lemma 2.5 is incorrect as written. The paper obtains ∫|h∘g_{k,m}-h∘g| dµ ≤ (2+2C)/k + 2C/(k−1) from Lemma 2.4, and then writes ∫|h∘g_{k,m}-h∘g|² dµ ≤ 2C∫|h∘g_{k,m}-h∘g| dµ ≤ 4C + 4C²/k + 4C²/(k−1). The last right-hand side tends to 4C as k→∞, not to 0, so the claim 'by arbitrariness of k>1 we have that h is almost periodic' does not follow from the displayed inequality. The intended inequality is (4C+4C²)/k + 4C²/(k−1), which does tend to 0. Since Lemma 2.5 is used in Theorem 2.7 for the converse direction of Theorem 2.1 (and hence of Theorem 1.1), the proof as printed has a genuine gap in a load-bearing step; however, the correction is immediate and does not affect the rest of the argument.","section":"Lemma 2.5 (Section 2)"}],"minor_comments":[{"comment":"There is a typographical error in the abstract: 'meas ures' should be 'measures'.","section":"Abstract"},{"comment":"The conclusion 'In particular, µ has bounded complexity w.r.t. {H_{F_n}}' is slightly too quick: the proof constructs sets on which H_{F_n}(x,y)<ε, whereas the complexity definition uses balls of radius ε/2. This is harmless because one can apply the statement with ε/2, but the sentence should say so.","section":"Theorem 2.6, proof of bounded complexity"},{"comment":"The sentence 'Without loss of generality, we may assume that the Følner sequence {F_n:n∈N} is tempered' is terse. A reader should be told that bounded complexity passes to any tempered subsequence (which exists by [12, Proposition 1.4]), and that the conclusion that h is almost periodic is independent of the choice of Følner sequence.","section":"Theorem 2.7, beginning of proof"},{"comment":"The statement says 'if and only if', but the proof only gives the direction '⇒'. The converse is indeed trivial by applying the stated property with ε/2 to obtain sets of H_{F_n}-diameter less than ε/2, but a one-sentence justification would improve readability.","section":"Lemma 2.2"}],"recommendation":"major_revision","confidential_remarks":"The paper is essentially correct in conception and execution; the only substantive issue is the arithmetic slip in Lemma 2.5, which is clearly fixable. I would be willing to accept the paper after the authors correct that estimate and clarify the two terse points listed in the minor comments."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline result is real: for countable amenable group actions, the paper proves the equivalence of discrete spectrum, bounded measure complexity (metric and Hamming), and, along tempered Følner sequences, mean equicontinuity and equicontinuity in the mean. Prior work covered Z-actions and abelian groups, so this is the right next step and the proofs are genuinely new. Section 2 does the heavy lifting: bounded complexity implies almost periodicity via Lindenstrauss's pointwise ergodic theorem, and the converse via finite-dimensional invariant subspaces. I checked the counting arguments in Theorem 2.7 and the measure estimates in Lemma 2.4; they are internally consistent.\n\nSoft spots: one genuine flaw in the written proof of Lemma 2.5. The displayed bound reads ≤ 4C + 4C^2/k + 4C^2/(k-1), which tends to 4C, not 0, so the conclusion “by arbitrariness of k > 1” does not follow as written. The intended correction is immediate: multiplying Lemma 2.4's bound by 2C gives (4C+4C^2)/k + 4C^2/(k-1), which does tend to 0. So this is a typo, not a conceptual hole, but it sits in a lemma used for the converse direction of the main equivalence, and a referee should require the fix. Two smaller issues: the abstract states the equicontinuity equivalences without the tempered hypothesis, which the theorem correctly includes; and the reduction of non-tempered Følner sequences to tempered subsequences in Theorem 3.1 is compressed, though the step is standard. No circularity: the central claims are derived from Zimmer's characterization of discrete spectrum and Lindenstrauss's pointwise theorem, with the first author's Z-action paper cited only for context.\n\nBottom line: this is a solid, citable contribution for ergodic theorists working on amenable group actions, and it deserves a serious referee. I would send it to peer review and ask for a corrected Lemma 2.5 rather than desk-rejecting.","headline":"The paper completes the discrete-spectrum/bounded-complexity/mean-equicontinuity package for countable amenable group actions with proofs that mostly hold up; one arithmetic typo in Lemma 2.5 needs a trivial fix.","tokens_in":17435,"tokens_out":2325,"would_cite":true,"duration_ms":21429,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37A35","37A15"],"pacs":[],"model":"deepseek-v4-flash","headline":"For amenable group actions, discrete spectrum is exactly bounded complexity, and for tempered Følner sequences exactly mean equicontinuity.","keywords":["discrete spectrum","amenable group actions","almost periodic functions","bounded complexity","mean equicontinuity","equicontinuity in the mean","measure complexity","Hamming distance"],"falsifier":"Take a nontrivial Bernoulli shift action of a countable amenable group with its product measure, choose any Følner sequence, and compute the minimal covering number $C(d_{F_n},\\varepsilon)$ for the metric averages. The theorem predicts this number is unbounded in $n$ for every small $\\varepsilon$, because the measure is not discrete spectrum; finding a Følner sequence along which it stays bounded would refute the implication from bounded complexity to discrete spectrum.","tokens_in":16243,"feed_emoji":"📐","tokens_out":14168,"duration_ms":117882,"temperature":0.7,"pith_summary":"The paper establishes a single characterization of discrete spectrum for invariant measures of countable amenable group actions on compact metric spaces. It proves that a measure has discrete spectrum exactly when its complexity with respect to any Følner-averaged metric is bounded, and exactly when the Hamming-distance complexity of partition names is bounded for every finite partition, even every two-element partition. For tempered Følner sequences, the same property is shown to be equivalent to mean equicontinuity and to equicontinuity in the mean. The relevance is that discrete spectrum is thereby not only an operator-theoretic notion but also a combinatorial rate-of-growth property, and this equivalence, previously known for integer actions, now holds for all countable amenable groups.","feed_headline":"Discrete spectrum equals bounded complexity for amenable group actions","feed_subtitle":"Spectral, complexity, and mean-equicontinuity descriptions all coincide, for every countable amenable group.","key_machinery":"The central object is the complexity function $C(w_{F_n},\\varepsilon)$, which counts the minimum number of balls of radius $\\varepsilon/2$ in the Følner-averaged semimetric $w_{F_n}(x,y)=\\frac{1}{|F_n|}\\sum_{g\\in F_n}w(gx,gy)$ needed to cover $X$ up to $\\mu$-measure $1-\\varepsilon$; bounded complexity means this count stays uniformly bounded in $n$ for each $\\varepsilon$. The companion object is the Hamming distance between partition names, $H_{\\alpha,F_n}(x,y)$, measuring the fraction of group elements in $F_n$ on which two points fall into different atoms of $\\alpha$. The load-bearing mechanism is a two-way passage: an almost periodic function can be approximated by finitely many finite-dimensional invariant subspaces, and the pointwise ergodic theorem for tempered Følner sequences turns that approximation into a finite uniform covering at every scale $\\varepsilon$; conversely, if complexity is bounded but the function is not almost periodic, averaging the Hamming distance over a Følner sequence produces many pairwise separated functions, contradicting the finite-dimensionality of the covering. This transfer is what makes statements (1)-(4) equivalent, and the same covering argument yields the mean-equicontinuity half.","core_discovery":"The central claim is Theorem 1.1: for a countable amenable group acting by homeomorphisms on a compact metric space and an invariant Borel probability measure µ, the following are equivalent for any Følner sequence $F_n$: (1) µ has discrete spectrum; (2) µ has bounded complexity with respect to the Følner-averaged metric averages $d_{F_n}$; (3) µ has bounded complexity with respect to Hamming-distance averages of names of every finite partition; (4) the same for every two-element partition. If the Følner sequence is tempered, these are also equivalent to (5) mean equicontinuity and (6) equicontinuity in the mean. The proof operates through the identification of discrete spectrum with almost periodicity of every square-integrable function, and through a two-way transfer between bounded complexity of a function's oscillation semimetric and almost periodicity. The result lifts the earlier equivalence for integer actions to all countable amenable groups.","pith_inferences":["The two-element partition criterion suggests a computable diagnostic: in a concrete simulation of a countable amenable group action, estimate the growth of Hamming-name complexity for a small collection of indicator functions; bounded growth across a Følner sequence is a numerical signature of discrete spectrum, while unbounded growth points to weak mixing or entropy.","The paper leaves open whether the mean-equicontinuity equivalence survives for non-tempered Følner sequences; constructing a non-tempered Følner sequence and a mean-equicontinuous measure without discrete spectrum, or proving the converse, would settle the boundary of the tempered hypothesis.","Because the proof only needs continuous semimetrics on compact spaces, the same equivalence plausibly extends to actions on compact Hausdorff spaces by replacing the metric with a separating family of continuous semimetrics, or to locally compact amenable groups with a suitable Følner averaging scheme."],"forward_implications":["An invariant measure of a countable amenable group action is spectrally pure precisely when its metric-averaged complexity is bounded, so discrete spectrum becomes a checkable rate-of-growth property independent of the group's internal structure.","Checking only two-element partitions suffices: if every measurable set has bounded Hamming-name complexity, then every finite partition does and the measure has discrete spectrum.","Under tempered Følner sequences, mean equicontinuity and equicontinuity in the mean are not merely related to discrete spectrum but identical to it; for such sequences all six viewpoints in Theorem 1.1 describe one property.","Every almost periodic function is individually characterized by bounded complexity of its own oscillation semimetric, so the technique applies function-by-function rather than only to the whole measure.","Because every Følner sequence has a tempered subsequence, the spectral/complexity equivalence holds for arbitrary Følner sequences, not only for the tempered ones where the pointwise ergodic theorem is available."],"supporting_citations":[{"why":"Supplies the pointwise ergodic theorem for tempered Følner sequences that powers Lemma 2.2 and Theorem 2.6.","marker":"[12]"},{"why":"Provides the theorem identifying discrete spectrum with almost periodicity of every L² function, the working definition used in all proofs.","marker":"[23]"},{"why":"Establishes the bounded-complexity/mean-equicontinuity equivalence for Z-actions that this paper extends to amenable groups.","marker":"[7]"},{"why":"Introduces measure complexity and proves the discrete-spectrum-implies-bounded-complexity direction for Z-actions.","marker":"[8]"},{"why":"Introduces Hamming-distance partition-name complexity for ergodic Z-actions, the framework behind statements (3) and (4).","marker":"[1]"},{"why":"Extends the partition-name complexity characterization to non-ergodic Z-actions and supplies the theorems being generalized.","marker":"[22]"},{"why":"Provides the admissible-metric characterization of discrete spectrum whose methods inspire the converse direction, Theorem 2.7.","marker":"[19]"},{"why":"The Z-action theorem characterizing almost periodic functions via complexity that Section 2 generalizes to amenable groups.","marker":"[5]"}],"fun_headline_variants":["Amenable group actions: discrete spectrum iff bounded complexity","Bounded complexity characterizes discrete spectrum for amenable groups","Discrete spectrum, bounded complexity, and mean equicontinuity unify for amenable groups","Amenable groups: discrete spectrum = bounded complexity = mean equicontinuity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the standard characterization of discrete spectrum as the condition that every square-integrable function has a small, finitely approximable orbit, together with Lindenstrauss's pointwise ergodic theorem for averages over tempered Følner sequences, is available; if either imported result gives way, the argument cannot pass from bounded complexity to discrete spectrum.","fun_headline_variants_meta":{"raw":{"variants":["Amenable group actions: discrete spectrum iff bounded complexity","Bounded complexity characterizes discrete spectrum for amenable groups","Discrete spectrum, bounded complexity, and mean equicontinuity unify for amenable groups","Amenable groups: discrete spectrum = bounded complexity = mean equicontinuity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00086,"raw_usage":{"total_tokens":3660,"prompt_tokens":804,"completion_tokens":2856,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":420,"completion_tokens_details":{"reasoning_tokens":2779}},"tokens_in":420,"tokens_out":2856,"duration_ms":19548,"temperature":1.0,"reasoning_tokens":2779,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:42:40.301899+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a nontrivial Bernoulli shift action of a countable amenable group with its product measure, choose any Følner sequence, and compute the minimal covering number $C(d_{F_n},\\varepsilon)$ for the metric averages. The theorem predicts this number is unbounded in $n$ for every small $\\varepsilon$, because the measure is not discrete spectrum; finding a Følner sequence along which it stays bounded would refute the implication from bounded complexity to discrete spectrum.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the pointwise ergodic theorem for tempered Følner sequences that powers Lemma 2.2 and Theorem 2.6."},{"cited_title":"Zimmer, Ergodic actions with generalized discrete spectrum , Illinois J","cited_arxiv_id":null,"evidence_quote":"Provides the theorem identifying discrete spectrum with almost periodicity of every L² function, the working definition used in all proofs."},{"cited_title":"Bounded complexity, mean equicontinuity and discrete spectrum","cited_arxiv_id":"1806.02980","evidence_quote":"Establishes the bounded-complexity/mean-equicontinuity equivalence for Z-actions that this paper extends to amenable groups."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces measure complexity and proves the discrete-spectrum-implies-bounded-complexity direction for Z-actions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces Hamming-distance partition-name complexity for ergodic Z-actions, the framework behind statements (3) and (4)."},{"cited_title":"Measure-theoretic mean equicontinuity and bounded complexity","cited_arxiv_id":"1807.05868","evidence_quote":"Extends the partition-name complexity characterization to non-ergodic Z-actions and supplies the theorems being generalized."},{"cited_title":"V ershik, Pavel B","cited_arxiv_id":null,"evidence_quote":"Provides the admissible-metric characterization of discrete spectrum whose methods inspire the converse direction, Theorem 2.7."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The Z-action theorem characterizing almost periodic functions via complexity that Section 2 generalizes to amenable groups."}],"review_version":1}