{"id":"9c86d637-5d09-4d17-be91-8b34df9462eb","arxiv_id":"2411.13519","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"An explicit, choice-free construction gives an uncountable proper subring of the real numbers with Hausdorff dimension zero.","lead":"This paper explicitly constructs an uncountable set of real numbers that is closed under addition, subtraction, and multiplication, yet is so sparse that its Hausdorff dimension is zero. The construction uses sums of powers of two with bounded digit coefficients and avoids the Axiom of Choice and the Continuum Hypothesis.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Ring closure hinges on Lemma 2's uniform representation bound; the proof's missing c_{n-1}(0) edge case is repairable, so the central claim survives.","rationale":"The central construction is sound. The set A is an increasing union of subgroups A_n, each of Hausdorff dimension zero because A_{n,t} is covered by 2^{-ℓ}-small intervals whose count grows only polylog in ℓ. The only place where the argument could genuinely fail is the multiplication closure: if c_n(k) grew with k, the coefficients of a product could be unbounded and the product would not lie in any A_{m+n}. Lemma 2 is therefore the load-bearing lemma. Its proof is correct in substance: the largest summand in any representation of k has exponent in an interval of size about log_2 n, and the induction hypothesis bounds the remaining n-1 terms. The sole omission is the treatment of k-2^m=0, which requires c_{n-1}(0); this is not a counterexample to the lemma but a missing case in the written induction. The Borel claim, though asserted without proof, is immediate from compactness of A_{n,t}. Thus the main theorem—an explicit uncountable proper subring of R with Hausdorff dimension zero—is established modulo a two-line repair. The reader's CONDITIONAL verdict is the right calibration.","tokens_in":4737,"tokens_out":42971,"duration_ms":444780,"concrete_test":"Re-derive Lemma 2 with the induction stated for all k∈N_0, explicitly defining c_n(0)=1 and requiring b_n≥1; then check that the largest-summand step covers k-2^m=0 and that the resulting bound b_n = n(1+log_2 n)b_{n-1} is finite for each fixed n. If this re-derivation goes through, the edge-case gap is closed and the ring-closure proof is complete.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing step is Lemma 2, which bounds c_n(k), the number of ordered representations of k as a sum of n elements of S={0,1,2,4,...}. Only through this bound does Theorem 3 conclude that the coefficients z_r of a product remain uniformly bounded, so xy ∈ A_{m+n}. The proof picks a summand ≥ k/n, placing its exponent in [log_2 k - log_2 n, log_2 k], and uses induction on c_{n-1}(k-2^m). As written, the lemma states c_n(k) for k∈N, but the edge case k-2^m=0 requires c_{n-1}(0), which is not defined; this is a genuine gap in the written proof. It is trivially repaired by defining c_{n-1}(0)=1 (the all-zero representation) and taking b_{n-1}≥1, and the interval bound #In,k≤1+log_2 n then yields the same finite b_n. The separate Borel assertion is unproved but follows because A_{n,t} is the continuous image of the compact product [-t,t]^{nS}, hence compact, and A is sigma-compact. Neither issue undermines the existence of an uncountable proper subring of Hausdorff dimension zero.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs an explicit set A = ∪_{n≥1} A_n, where A_n consists of real numbers representable by a base-2-type series with bounded integer coefficients supported on the set nS, the n-fold sumset of S = {0,1,2,4,8,...}. The authors prove that each A_n is an uncountable subgroup of R, that each truncated stage A_{n,t} has Hausdorff dimension zero by an explicit interval-covering argument, and hence that A has Hausdorff dimension zero. They then prove that A is closed under multiplication using a uniform bound on the number of representations of an integer as a sum of n elements of S (Lemma 2), making A a subring. A final argument shows that the only rational numbers in A are the dyadic rationals, so A is a proper subring and is not a field, and the paper also asserts that A is Borel. The main claim of the manuscript is that A is an uncountable proper subring of R with Hausdorff dimension zero.","tokens_in":5011,"tokens_out":20181,"duration_ms":220262,"significance":"If the small gaps noted below are repaired, this is a short, explicit, parameter-free construction of an uncountable proper subring of R with Hausdorff dimension zero, complementing the known results of Erdős–Volkmann (subgroups of any Hausdorff dimension), Davies (CH-dependent non-Borel subrings), and Edgar–Miller (Borel subrings have dimension zero). The proof is self-contained and verifies the ring and dimension properties directly from the definitions, with Lemma 2 as the key technical estimate. The construction uses no free parameters and no choice principle, and the main claims are falsifiable in the sense that each step is checked against the definitions of Hausdorff measure, subgroups, and rings. Once the Borel assertion is supplied with a proof, the paper also provides an explicit Borel example, which is a useful complement to the non-Borel examples in the literature.","major_comments":[],"minor_comments":[{"comment":"The induction step uses c_{n-1}(k-2^m) even when k-2^m = 0, although c_{n-1} is defined only for positive integers in the lemma statement. This is a small but real gap in a lemma that underpins Theorem 3; it should be repaired, for example by defining c_j(0) = 1 (the all-zero representation) and requiring b_j ≥ 1, so that the estimate c_n(k) ≤ n b_{n-1}(1 + log_2 n) remains valid.","section":"Lemma 2"},{"comment":"The opening of Section 2 states that the constructed ring is a Borel set, but no proof is given. This follows because each A_{n,t} is the continuous image of the compact product [-t,t]^{nS} and hence compact, so A is sigma-compact; the authors should include this argument or explicitly mark the Borel property as a separate claim.","section":"Section 2, Borel claim"},{"comment":"The uncountability argument invokes uniqueness of binary expansions, but the sums are supported on the sparse set nS rather than on all of N0. The claim is true (at the first differing position, the tail over nS is strictly smaller than the digit gap because nS omits infinitely many integer positions), but a one-sentence proof would make the argument fully rigorous.","section":"Theorem 1"},{"comment":"In the proof of Theorem 3, the sentence 'for every k ∈ nS there exist x_k ∈ [-t,t]_Z and y_k ∈ [-t',t']_Z' should refer to k ∈ mS for the x_k coefficients; as written the index set is inconsistent with the displayed sums.","section":"Theorem 3"},{"comment":"The line 'ℓ ≥ 2c_l^r − 2 where r = 1/n' appears to be a typographical corruption of ℓ ≥ 2^{c_l^{1/n} - 2}, which is what the preceding inequality c_l ≤ (2 + log_2 ℓ)^n actually yields; please correct the exponent.","section":"Theorem 2, proof of Eq. (2)"},{"comment":"The proof only establishes that a rational element of A has a finite binary expansion; the converse, that every dyadic rational lies in A, should be stated (it follows from 1/2^k ∈ A_1 for k ∈ S and closure under addition).","section":"Corollary 1"},{"comment":"There are several typographical errors: 'A is a a subring' in Theorem 3, the inconsistent spelling of Erdős, and 'for every z ∈ N' in Lemma 1 should be 'for every n ∈ N'.","section":"Throughout"}],"recommendation":"minor_revision","confidential_remarks":"The central construction is sound and the result is interesting as a short, explicit counterpoint to the CH-dependent non-Borel subring constructions. The missing Borel proof and the small gap in Lemma 2 are easily repairable and do not affect the main theorem. I see no concerns about novelty or attribution; the citations to [1], [2], and [4] are appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper delivers exactly what the title says, and the main proof is sound. It is the first fully explicit, choice-free, Borel example of an uncountable subring of R with Hausdorff dimension zero. That is worth having.\n\nWhat is genuinely new: Erdős–Volkmann built subgroups of any dimension; Edgar–Miller proved every proper Borel subring has dimension zero but did not construct an uncountable one; Davies's example uses CH and is non-Borel. Here the series-expansion construction is concrete, and the key step is Lemma 2, the uniform bound on representations of integers as sums of powers of two. That bound is what makes multiplication close: without it the product coefficients could grow with r. The proof of Lemma 2 is clear and correct, up to one unstated edge case. The covering argument for dimension zero is standard but clean, and the dyadic-rational observation (Corollary 1) is a nice extra.\n\nSoft spots, both minor. First, the abstract and Section 2 say the ring is Borel, but no proof is given. It is easily supplied: each A_{n,t} is the continuous image of the compact product [-t,t]^{nS} (with the product topology), hence compact, so A is sigma-compact. The authors should add a sentence. Second, Lemma 2 invokes c_{n-1}(k-2^m) even when k-2^m=0; c_{n-1}(0) is not defined by the lemma's statement. The fix is to define c_{n-1}(0)=1 and note b_{n-1}≥1 works. Neither gap touches the main existence claim. There is also a harmless typo in Theorem 3 (\"a a subring\").\n\nThe citation pattern is honest: [1],[2],[4] are context and standard facts, and the paper openly says the oral construction it mentions comes without a reference. That is transparent rather than a defect.\n\nWho is this for? Anyone working on additive combinatorics or fractal geometry in the reals; also descriptive set theorists interested in what can be done without CH. It is a clean example, not a new research program, and it does not generalize dramatically. But it is a solid, citable construction.\n\nI would send it to a serious referee. The two gaps are easily repairable and do not undermine the central claim.","headline":"Explicit, choice-free Borel subring of R with Hausdorff dimension zero; the construction is sound and the two soft spots are minor and repairable.","tokens_in":5509,"tokens_out":1996,"would_cite":true,"duration_ms":18860,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["28A78","11K55","11B13"],"pacs":[],"model":"deepseek-v4-flash","headline":"An uncountable subring of $\\mathbb{R}$ with Hausdorff dimension zero","keywords":["Hausdorff dimension","subring of the reals","binary expansion","sumsets of powers of two","Borel set","Lebesgue measure zero","dyadic rationals","additive subgroup"],"falsifier":"Search for integers k whose number of ordered representations as a sum of n elements of S is unbounded as k grows; any unbounded sequence would falsify Lemma 2 and break the multiplication-closure argument. More directly, exhibit two elements of A whose product, expanded at positions in any (m+n)S, has coefficient sequence not bounded by any fixed integer.","tokens_in":1777,"feed_emoji":"📏","tokens_out":2219,"duration_ms":137920,"temperature":0.7,"pith_summary":"This paper constructs a subset $A$ of the real numbers that is an uncountable proper subring, yet is as thin as possible geometrically: its Hausdorff dimension is zero, so in particular it has Lebesgue measure zero. The construction is explicit, using binary-style series whose nonzero digit positions are restricted to sumsets of powers of two. Each intermediate set $A_n$ is an uncountable additive subgroup of dimension zero, and the union $A$ is closed under multiplication. The paper also shows the only rational numbers in $A$ are dyadic rationals, so $A$ is not a field.","feed_headline":"Uncountable ring of reals built with Hausdorff dimension zero","feed_subtitle":"Sparse binary-style expansions put uncountably many reals into a set with Lebesgue measure zero.","key_machinery":"The central object is the sparse digit-position set $S=\\{0\\}\\cup\\{2^m:m\\ge0\\}$ together with its iterated sumsets $nS$. Elements of $A_n$ are series $\\sum_{k\\in nS} x_k 2^{-k}$ with integer coefficients bounded by some $t$. The sparseness of $nS$ drives the dimension computation: the number of allowed positions up to $\\ell$ is at most $(2+\\log_2\\ell)^n$, so the natural truncation covers shrink fast enough to force the $s$-dimensional Hausdorff measure to vanish for every $s>0$. Multiplicative closure is carried by Lemma 2, a uniform bound on the ordered representation count $c_n(k)$, the number of ways to write $k$ as a sum of $n$ elements of $S$. That bound transfers a product of two bounded-digit series into another series of the same type with bounded digits, placing the product in $A_{m+n}$.","core_discovery":"Let $S=\\{0\\}\\cup\\{2^m:m\\ge 0\\}$ and let $nS$ be the set of all sums of $n$ elements of $S$. The paper defines $A_n$ as the set of numbers of the form $\\sum_{k\\in nS} x_k 2^{-k}$ with integer coefficients bounded in absolute value by some $t$. It proves that each $A_n$ is an uncountable subgroup of $\\mathbb{R}$ (Theorem 1) and that $A_n$ has Hausdorff dimension zero (Theorem 2). Letting $A=\\bigcup_{n\\ge 1} A_n$, the paper proves $A$ is closed under multiplication (Theorem 3) using a uniform bound on the number of representations of an integer as a sum of $n$ elements of $S$ (Lemma 2). Hence $A$ is an uncountable proper Borel subring of $\\mathbb{R}$ of Hausdorff dimension zero; Corollary 1 shows its rational elements are exactly the dyadic rationals, so $A$ is not a field.","pith_inferences":["The paper does not pursue this, but the same construction with $S$ replaced by powers of any integer $b\\ge3$, or by other sparse sets satisfying the logarithmic-density condition, should produce further explicit zero-dimensional subrings.","Because Corollary 1 excludes non-dyadic rationals, the method cannot directly produce a subfield; an uncountable Borel subfield of $\\mathbb{R}$ of dimension zero, if one exists, would need a different digit set or extra closure machinery.","The explicit covers suggest a computational check: for fixed $n$ and $t$, one can enumerate the truncated expansions and numerically verify that the predicted $\\delta$-covers shrink as claimed, giving an independent confirmation of the dimension-zero bound in small cases."],"forward_implications":["There is an explicitly defined uncountable proper Borel subring of $\\mathbb{R}$ with Hausdorff dimension zero, so the phenomenon of large algebraic structures inside geometrically negligible sets is not limited to additive subgroups.","Every rational element of $A$ has the form $a/2^k$, so $A$ contains no rational like $1/3$ and is therefore not a field.","Each element of $A$ has binary expansions containing arbitrarily long runs of equal digits (Proposition 1), a concrete Diophantine property that can be studied independently.","The construction goes through for any digit set $T\\cup\\{0\\}$ with bounded logarithmic density (Remark 1), yielding a whole family of zero-dimensional subrings rather than a single example.","The ring $A$ is a nested union of uncountable additive subgroups $A_n$ of dimension zero, so the step from subgroups to a subring is exactly where the uniform representation-count bound enters."],"supporting_citations":[{"why":"Shows that every proper Borel subring of $\\mathbb{R}$ has Hausdorff dimension zero; the paper's example is an explicit instance of this phenomenon.","marker":"[1]"},{"why":"Constructs Borel additive subgroups of $\\mathbb{R}$ with prescribed Hausdorff dimension; the paper compares its series construction with it and uses it to frame the open subring question.","marker":"[2]"},{"why":"Records the earlier unpublished construction of subrings of $\\mathbb{R}$ of arbitrary Hausdorff dimension under the Continuum Hypothesis, which the present paper replaces with an explicit, choice-free construction.","marker":"[3]"},{"why":"Supplies the definition of Hausdorff measure and dimension, and the countable-union property used to conclude that the union over n of the sets $A_n$ has dimension zero.","marker":"[4]"}],"fun_headline_variants":["Uncountable subring of R with zero Hausdorff dimension","Sparse binary sums build uncountable ring of measure zero","Zero-dimension uncountable ring: a Borel subring of R","Uncountable ring inside R, yet Hausdorff dimension zero","Sparse binary expansions yield uncountable zero-dimension ring"],"cache_read_input_tokens":7680,"weakest_assumption_plain":"The proof that products stay inside the ring depends on the claim that for each n there is a uniform upper bound, independent of the integer k, on the number of ways to write k as a sum of n numbers from {0,1,2,4,8,...}; if that bound failed, product coefficients could grow without bound and the ring could spill outside the construction.","fun_headline_variants_meta":{"raw":{"variants":["Uncountable subring of R with zero Hausdorff dimension","Sparse binary sums build uncountable ring of measure zero","Zero-dimension uncountable ring: a Borel subring of R","Uncountable ring inside R, yet Hausdorff dimension zero","Sparse binary expansions yield uncountable zero-dimension ring"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000886,"raw_usage":{"total_tokens":3738,"prompt_tokens":770,"completion_tokens":2968,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":386,"completion_tokens_details":{"reasoning_tokens":2880}},"tokens_in":386,"tokens_out":2968,"duration_ms":21419,"temperature":1.0,"reasoning_tokens":2880,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T16:20:26.027738+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Search for integers k whose number of ordered representations as a sum of n elements of S is unbounded as k grows; any unbounded sequence would falsify Lemma 2 and break the multiplication-closure argument. More directly, exhibit two elements of A whose product, expanded at positions in any (m+n)S, has coefficient sequence not bounded by any fixed integer.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Shows that every proper Borel subring of $\\mathbb{R}$ has Hausdorff dimension zero; the paper's example is an explicit instance of this phenomenon."},{"cited_title":"Erd˝ os and B","cited_arxiv_id":null,"evidence_quote":"Constructs Borel additive subgroups of $\\mathbb{R}$ with prescribed Hausdorff dimension; the paper compares its series construction with it and uses it to frame the open subring question."},{"cited_title":"Falconer, On the Hausdorff dimensions of distance sets, Mathematika 32 (1985), 206–212","cited_arxiv_id":null,"evidence_quote":"Records the earlier unpublished construction of subrings of $\\mathbb{R}$ of arbitrary Hausdorff dimension under the Continuum Hypothesis, which the present paper replaces with an explicit, choice-free construction."},{"cited_title":"Falconer, Fractal geometry, Mathematical foundations and applications(Second Edition), John Wiley & Sons Inc., Hoboken, NJ, 2003","cited_arxiv_id":null,"evidence_quote":"Supplies the definition of Hausdorff measure and dimension, and the countable-union property used to conclude that the union over n of the sets $A_n$ has dimension zero."}],"review_version":1}