{"id":"a0b5319e-055c-436e-867a-1114d55bf0da","arxiv_id":"2608.14395","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A central Cantor set is the range of exactly one purely atomic measure iff its fast-convergent sequence satisfies x_n ≠ 2x_{n+1}; for symmetric Cantor sets, iff the representation is irreducible with multiplicities at most two.","lead":"The paper identifies exactly when a Cantor set of possible subset sums is the range of one and only one atomic measure: for central Cantor sets, no term of the representing sequence may be twice the next. This settles a recovery question from measure theory and adds new examples of uniquely determined Cantorval ranges.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unproved 'Observation' in Lemma 6 is load-bearing for Theorem 5 and is also miscounted as written (M=∏M_i cannot index the E-gaps).","rationale":"The paper's main new result is Theorem 5, and its 'if' direction rests entirely on Lemma 6. The reader identified the unproved Observation inside Lemma 6 as the weakest assumption; my reading agrees, and I find the concern sharper than a mere missing justification: as reproduced, the Observation's indexing M=∏M_i does not match the number of E-gaps of order at most n. This makes the proof of (10) unverifiable as written. The central Cantor Theorem 4 is supported by the Banakiewicz theorem and a self-contained argument, and Theorem 8's proof appears valid under its stated hypotheses, so a wholesale rejection is not warranted. However, the symmetric Cantor characterization should not be accepted unconditionally until the Observation is either proved, corrected, or replaced. This keeps the reader's CONDITIONAL verdict unchanged.","tokens_in":16361,"tokens_out":26758,"duration_ms":232238,"concrete_test":"Verify the Observation in a permitted case of Lemma 6: take M_1=M_2=1 with α_1=4, α_2=3, α_3=1/2, and a fast-convergent tail with sum less than 1/10 (e.g. α_4=0.05, α_5=0.02, α_6=0.01, ...). Enumerate all E-gaps of order at most 2 in natural order. There are three gaps, so M=∏M_i=1 does not index them; restate the Observation with the correct count ∏(M_i+1)−1 and test the adjacency claim that every shorter gap is followed by a gap of length α_2−R_2. If the corrected claim fails, or if it cannot be derived from the First and Third Gap Lemmas, the proof of (10) needs a separate argument before Theorem 5 is accepted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 6 is the engine of the 'if' direction of Theorem 5, and its proof of (10) hinges on the unproved 'Observation'. The Observation is invoked exactly where the argument must force the next pseudointerval to align with a long gap: after a short gap between P_d^(1) and P_{d+1}^(1), it asserts that the gap between P_{d+1}^(1) and P_{d+2}^(1) has length exactly α_n−R_n, and this alignment then drives the contradiction in Case 1. No proof or citation is supplied. There is also a more elementary difficulty: under the paper's own definition, the number of E-gaps of order at most n in E(α_p;M_p) is ∏_{i=1}^n(M_i+1)−1, not M=∏_{i=1}^n M_i. For the central Cantor case M_i≡1, n=2, there are three such gaps while M=1. Thus the Observation is either misstated or depends on a non-obvious special meaning of 'order at most n'; in either reading, the gap-alignment argument in the proof of (10) is not currently supported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies unique achievability of achievement sets of summable positive non-increasing sequences, interpreted as ranges of purely atomic finite measures. The main results are: Theorem 4 characterizes uniquely achievable central Cantor sets by the condition x_n ≠ 2x_{n+1} for all n; Theorem 5 characterizes uniquely achievable symmetric Cantor sets as precisely those with an irreducible semi-fast representation (α_n; M_n) and M_n ≤ 2 for all n; and Theorem 8 gives a sufficient condition for unique achievability of a class of Cantorvals, yielding new examples beyond the Guthrie–Nymann Cantorval. The proof of Theorem 4 uses the Banakiewicz theorem on centers of distances, while Theorem 5 relies on a detailed geometric analysis of pseudointervals and a long Lemma 6. The paper closes with remarks and open problems on unique achievability of Cantorvals.","tokens_in":16670,"tokens_out":19771,"duration_ms":160578,"significance":"If correct, Theorem 4 is a clean complete characterization for central Cantor sets, and Theorem 5 is a substantial extension to the broader class of symmetric Cantor sets. Theorem 8 is a short, checkable argument that produces the first new uniquely achievable Cantorvals beyond the Guthrie–Nymann example. The center-of-distances argument in Theorem 4 is elegant, and the statement of Theorem 8 is falsifiable and clearly illustrated by the multigeometric examples. The main risk is in the proof of Theorem 5, which depends on Lemma 6 and on an induction over dominating gaps; several steps in that proof are not currently fully supported.","major_comments":[{"comment":"The Observation as stated cannot be correct. With M := ∏_{i=1}^n M_i, the list (L_j)_{j=1}^M cannot enumerate all E(y_i)-gaps of order at most n: for the central Cantor case M_i ≡ 1 and n = 2, there are three gaps of order at most 2 (one of order 1 and two of order 2), while M = 1. In general the number of such gaps is ∏_{i=1}^n(M_i+1) − 1. Moreover, the Observation is asserted without proof. It is used in two decisive places in the proof of (10), namely the case d ≤ s−2 and the induction step for the pseudointervals W_t, to conclude that the next pseudointerval gap has length exactly α_n − R_n. Until the Observation is corrected and proved, the alignment argument that establishes (10) is unsupported.","section":"Section 2, Lemma 6, 'Observation'"},{"comment":"The proof begins with the assertion 'Since E has infinitely many gaps, it has infinitely many dominating gaps.' This contradicts the definition of a dominating gap given in Section 1 (all E-gaps lying to the left are shorter). In a compact set, the lengths of gaps tend to 0 near the right endpoint, so there are only finitely many record highs from the left; for example, the ternary Cantor set E(1/3^n), which satisfies the hypotheses of Theorem 5, has a unique dominating gap under this definition. The induction over the sequence (n_i) therefore needs a different construction: presumably the indices should be chosen recursively as the longest dominating gap of each successive tail E∩[0,R_{n_i}]. This must be stated and proved before the induction is valid.","section":"Section 2, proof of Theorem 5 (⇐)"},{"comment":"Several steps in Case 2 are deferred to 'an argument analogous to the proof of (10)' and 'in a manner analogous to the final part of the proof that (6) and (7) cannot hold simultaneously'. Since the proof of (10) itself is not yet established (see the comments on the Observation), and since the arguments rely on visual inspection of Figures 1–3, this part of the case analysis is not sufficiently self-contained. The authors should either give a complete proof of (10) and then spell out the analogous arguments for Case 2, or restructure the proof so that every alignment of pseudointervals is justified by an explicit combinatorial statement rather than by reference to figures.","section":"Section 2, proof of Lemma 6, Case 2"}],"minor_comments":[{"comment":"There are numerous typographical and OCR artifacts, for example 'PUREL Y A TOMIC' in the title, 'Gł¸ ab' in the references, and 'j∈s+1,...,s,' in Case 2 of Lemma 6. These should be cleaned up before publication.","section":"Throughout"},{"comment":"In the proof of (10), the sentence 'If min ˜P2 ∈ [min P^(2)_1, max P^(2)_1], then max ˜P2 ∈ P^(2)_2' does not address the case where min ˜P2 lies in the gap (max P^(2)_1, min P^(2)_2); this intermediate case appears to be excluded only implicitly, and the exclusion should be justified.","section":"Section 2, proof of Lemma 6"},{"comment":"The proof that K ≠ {2n} forces the existence of m with m, m+1 ∉ K and m+2 ∈ K is stated in one sentence and deserves a short justification, since the preceding dichotomy only rules out consecutive elements of K and three consecutive non-elements of K.","section":"Section 3, Proposition 10"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the main claims are interesting. The central concern is the completeness of Lemma 6 and the dominating-gap induction in Theorem 5; these are fixable, but they are load-bearing. If the authors can supply a correct proof of the Observation and clarify the induction over successive tails, the paper would likely be acceptable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The central Cantor result is the real content here, and it is good. Theorem 4 gives a clean necessary and sufficient condition (x_n ≠ 2x_{n+1}) for unique achievability, proved with Banakiewicz's center-of-distances theorem plus a short minimal-center argument. That part is checkable and, as far as I can tell, correct. The symmetric Cantor theorem and the Cantorval section are more speculative.\n\nWhat's new: beyond Theorem 4, the paper extends to symmetric Cantor sets (Theorem 5) and gives a sufficient condition for Cantorval uniqueness (Theorem 8), which yields genuinely new examples. The proof of Theorem 8 is clean and reproducible. The citation pattern is fine: heavy use of the same circle's results, but those are published, and Theorem 4 is not a restatement of them.\n\nSoft spots: The 'if' direction of Theorem 5 hangs on Lemma 6, and Lemma 6 hangs on an unproved 'Observation' about the ordering of E-gaps of order at most n. The Observation is load-bearing: it is exactly what forces the alignment of pseudointervals in the proof of (10), and without it Case 1 collapses. Worse, the Observation as written is miscounted. It says M = ∏_{i=1}^n M_i gaps, but by the paper's own definition of gap order, the number of E-gaps of order at most n is ∏(M_i+1)−1. For the central Cantor case M_i=1, n=2, that's three gaps, not one. I don't see an interpretation that saves the statement; if the Observation is just a typo, the proof still needs the underlying gap-length claim, which is not proven or cited. The long case analysis in Lemma 6 also leans on several 'clearly' steps that are only visible in figures.\n\nProportion: These are real issues, but not necessarily fatal. Theorem 4 stands on its own. Theorem 5 may be true; the necessity side is proven by explicit rearrangement, and the sufficiency side is a proof gap rather than a known counterexample. The paper reads like a serious working note, not a finished paper.\n\nWho it's for: people in achievement set theory and anyone working on Banakh's measure-recovery question. A referee should be sent to it, but the referee should be told to focus on Lemma 6. If the authors can prove the Observation (or replace it with a valid argument), the paper is a solid contribution.","headline":"A genuinely useful central Cantor characterization, with a symmetric-set extension that is conditional on an unproved and miscounted gap-order observation.","tokens_in":17131,"tokens_out":5168,"would_cite":true,"duration_ms":43054,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":["40A05","11B05","28A75"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper characterizes when a purely atomic finite measure is uniquely determined by its range: for central Cantor sets no canonical term may be twice its successor, and for symmetric Cantor sets the representation must be irreducible…","keywords":["atomic measure","range of a measure","achievement set","central Cantor set","symmetric Cantor set","Cantorval","center of distances","unique achievability"],"falsifier":"A concrete check: take a symmetric Cantor set satisfying irreducibility and $M_n \\leq 2$ (for example multiplicities $M=(2,1,1,\\ldots)$ with $\\alpha_p$ chosen so that no $\\alpha_p$ equals $(M_{p+1}+1)\\alpha_{p+1}$), list the $E$-gaps of order at most $n$ in increasing order, and look for a gap shorter than $\\alpha_n-R_n$ that is either the last such gap or is followed by a gap of a different length. One such example would refute the unproved Observation and break the supplied proof of the direction of Theorem 5 asserting that irreducibility with multiplicities at most two forces uniqueness.","tokens_in":16184,"feed_emoji":"🧮","tokens_out":13423,"duration_ms":108221,"temperature":0.7,"pith_summary":"This paper settles, for two large families of self-similar ranges, when a finite purely atomic measure is the only measure with that range. The range of such a measure is the achievement set of its atom weights, the set of all subsums, and the question is whether that set determines the weights uniquely. For central Cantor sets with left endpoint $0$, uniqueness is equivalent to a single local condition on the canonical fast-convergent representation $(x_n)$: no term equals twice the next term. For symmetric Cantor sets, uniqueness is equivalent to the semi-fast representation $E(\\alpha_n; M_n)$ being irreducible, with $\\alpha_n \\neq (M_{n+1}+1)\\alpha_{n+1}$ for all $n$, and satisfying $M_n \\leq 2$ for all $n$. The paper also proves a sufficient condition under which certain Cantorval ranges are uniquely achievable, producing new examples beyond the previously known one.","feed_headline":"Cantor sets pin down their measure unless a term doubles the next","feed_subtitle":"No doubled terms means one measure; reducible values and triple repeats spoil uniqueness.","key_machinery":"The central object is the achievement set $E(x_n) = \\{\\sum_{n\\in A} x_n : A \\subseteq \\mathbb{N}\\}$ of a summable non-increasing sequence; the range of the measure is exactly this set. The main diagnostic used for central Cantor sets is the center of distances $C(E)$, the set of all $\\alpha \\geq 0$ such that from every point of $E$ there is another point of $E$ at distance $\\alpha$. A known theorem characterizes when $C(E)$ is minimal for central Cantor sets in terms of three local ratio patterns, and that characterization is what forces $x_n = 2x_{n+1}$ in the non-unique case. For symmetric Cantor sets the proof mechanism is gap geometry: the longest dominating gap — a gap longer than every gap to its left — is principal, and the proof aligns any alternative representation against the canonical one by matching translated copies of remainder achievement sets, called pseudointervals, until the first block of terms coincides; an induction over the sequence of dominating gaps then forces all terms to coincide.","core_discovery":"The central discovery is a necessary-and-sufficient criterion. A central Cantor set with left endpoint $0$, in its unique fast-convergent representation $(x_n)$, is the range of exactly one purely atomic finite measure if and only if $x_n \\neq 2x_{n+1}$ for every $n$. If some $x_n = 2x_{n+1}$, the paper constructs an explicit different sequence with the same achievement set: the pair $(x_n,x_{n+1})$ is replaced by three copies of $x_{n+1}$. The converse is proved by passing through the center of distances: any alternative representation either forces a term to be repeated, which by the repeated-value lemma puts a multiple of that term into the center of distances and thereby produces the doubling, or makes the center of distances non-minimal, which by the known characterization also produces $x_n = 2x_{n+1}$. For symmetric Cantor sets the criterion is that the semi-fast representation $E(\\alpha_n; M_n)$ be irreducible, meaning $\\alpha_n \\neq (M_{n+1}+1)\\alpha_{n+1}$ for every $n$, and that every multiplicity satisfy $M_n \\leq 2$; the harder direction is proved by aligning pseudointervals from the longest dominating gap outward and inducting over all dominating gaps.","pith_inferences":["The results point to a local-ratio dichotomy for all achievement sets: non-uniqueness may always come from a regrouping of terms into a doubled, tripled, or reducible pattern. If so, the paper's Problem 12 would have an affirmative answer, but that is an extrapolation, not proved here.","The pseudointerval-matching argument in Lemma 6 is a rigidity mechanism that could transfer to other centrally symmetric or self-similar sets; a testable extension is to symmetric Cantor sets with non-uniform dissection ratios.","The unproved Observation about gap ordering is computationally checkable for the Cantorval families generated in the paper; a failure there would not necessarily refute Theorem 5 but would require a different proof."],"forward_implications":["For central Cantor sets, uniqueness of the range is equivalent to the local test $x_n \\neq 2x_{n+1}$ for every $n$; no other global condition is needed.","Every symmetric Cantor range failing the criterion is explicitly non-unique: reducible values admit a split-and-merge substitution, and any multiplicity at least $3$ admits a replacement representation.","Every uniquely achievable symmetric Cantor range has an irreducible semi-fast representation with no value occurring more than twice.","The sufficient condition of Theorem 8 yields uniquely achievable Cantorval ranges beyond the standard two-term Cantorval, including some sequences whose topological type is unknown.","In the Cantorval family generated by Theorem 9, the standard two-term Cantorval is the only uniquely achievable member; every other member contains a doubled term."],"supporting_citations":[{"why":"Supplies the theorem that the center of distances of a central Cantor set is non-minimal exactly under three local ratio patterns, used to force the doubling in Theorem 4.","marker":"[2]"},{"why":"Defines irreducible semi-fast sequences and the lemma that every element of the center of distances is a finite subsum, giving the key vocabulary and one rigidity fact for Theorem 5.","marker":"[3]"},{"why":"Provides the foundational facts on achievement sets, gap orders, the First Gap Lemma, and the representation of an achievement set as an intersection of iterates.","marker":"[6]"},{"why":"Establishes the correspondence between symmetric Cantor sets and semi-fast convergent sequences and the fact that their achievement sets are Cantor sets.","marker":"[7]"},{"why":"Supplies the recovery-from-range problem, the repeated-value lemma used in Theorem 4, and the earlier sufficient uniqueness conditions that these theorems complete.","marker":"[9]"},{"why":"Introduces the center of distances and the proposition that all terms of any representation lie in it, the main diagnostic behind Theorem 4.","marker":"[10]"},{"why":"Classifies achievement sets as Cantor sets or Cantorvals, fixing the class of ranges treated in the paper.","marker":"[17]"},{"why":"Provides the generator of Cantorval sequences used in Proposition 10 and in the examples illustrating Theorem 8.","marker":"[29]"}],"fun_headline_variants":["No doubled terms, one measure: Cantor set uniqueness","Doubled term forces second measure for Cantor sets","Cantor measure unique iff no term doubles the next","When Cantor sets fix a single measure: no doubling","Symmetric Cantor sets: irreducibility plus low multiplicities"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the unproved Observation in Lemma 6, which says that in the ordered list of gaps of order at most $n$ every gap shorter than $\\alpha_n-R_n$ is immediately followed by a gap of length exactly $\\alpha_n-R_n$; if that adjacency pattern fails, the pseudointerval-alignment proof of the uniqueness direction of Theorem 5 collapses.","fun_headline_variants_meta":{"raw":{"variants":["No doubled terms, one measure: Cantor set uniqueness","Doubled term forces second measure for Cantor sets","Cantor measure unique iff no term doubles the next","When Cantor sets fix a single measure: no doubling","Symmetric Cantor sets: irreducibility plus low multiplicities"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000331,"raw_usage":{"total_tokens":1803,"prompt_tokens":862,"completion_tokens":941,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":478,"completion_tokens_details":{"reasoning_tokens":860}},"tokens_in":478,"tokens_out":941,"duration_ms":8252,"temperature":1.0,"reasoning_tokens":860,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-27T21:13:32.697408+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete check: take a symmetric Cantor set satisfying irreducibility and $M_n \\leq 2$ (for example multiplicities $M=(2,1,1,\\ldots)$ with $\\alpha_p$ chosen so that no $\\alpha_p$ equals $(M_{p+1}+1)\\alpha_{p+1}$), list the $E$-gaps of order at most $n$ in increasing order, and look for a gap shorter than $\\alpha_n-R_n$ that is either the last such gap or is followed by a gap of a different length. One such example would refute the unproved Observation and break the supplied proof of the direction of Theorem 5 asserting that irreducibility with multiplicities at most two forces uniqueness.","supporting_citations":[{"cited_title":"Banakiewicz,The center of distances of central Cantor sets, Results Math.78(2023), art","cited_arxiv_id":null,"evidence_quote":"Supplies the theorem that the center of distances of a central Cantor set is non-minimal exactly under three local ratio patterns, used to force the doubling in Theorem 4."},{"cited_title":"Banakiewicz, A","cited_arxiv_id":null,"evidence_quote":"Defines irreducible semi-fast sequences and the lemma that every element of the center of distances is a finite subsum, giving the key vocabulary and one rigidity fact for Theorem 5."},{"cited_title":"Bartoszewicz, M","cited_arxiv_id":null,"evidence_quote":"Provides the foundational facts on achievement sets, gap orders, the First Gap Lemma, and the representation of an achievement set as an intersection of iterates."},{"cited_title":"Bartoszewicz, M","cited_arxiv_id":null,"evidence_quote":"Establishes the correspondence between symmetric Cantor sets and semi-fast convergent sequences and the fact that their achievement sets are Cantor sets."},{"cited_title":"Bartoszewicz, S","cited_arxiv_id":null,"evidence_quote":"Supplies the recovery-from-range problem, the repeated-value lemma used in Theorem 4, and the earlier sufficient uniqueness conditions that these theorems complete."},{"cited_title":"Bielas, S","cited_arxiv_id":null,"evidence_quote":"Introduces the center of distances and the proposition that all terms of any representation lie in it, the main diagnostic behind Theorem 4."},{"cited_title":"Guthrie, J.E","cited_arxiv_id":null,"evidence_quote":"Classifies achievement sets as Cantor sets or Cantorvals, fixing the class of ranges treated in the paper."},{"cited_title":"Nowakowski,On a new condition implying that an achievement set is a Cantorval and its applications, arXiv:2512.17761","cited_arxiv_id":null,"evidence_quote":"Provides the generator of Cantorval sequences used in Proposition 10 and in the examples illustrating Theorem 8."}],"review_version":1}