{"id":"919f1721-a6c9-4b58-83e9-98abb56fe358","arxiv_id":"2507.20867","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"high","formal_verification":"none","parameter_count":2,"one_line_summary":"A sigma-morphic protoset made entirely of convex polygons is constructed by replacing the bumps and nicks of a known non-convex example with angular convex notches.","lead":"This paper constructs the first known set of convex polygons that tiles the plane in exactly countably many non-congruent ways. It addresses an open variant of a 1977 problem about polymorphic tiles.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.2 is not established: the local vertex-forcing argument on which the whole construction rests is deferred ('we'll leave that for later') and never supplied, and the genericity claim for the parameters is asserted, not proved.","rationale":"The paper's central claim is a constructive existence theorem; to establish it one must prove both that the convex prototiles exist and that their tiling space is exactly countable. The manuscript gives a plausible route via Schmitt's protoset but not the proof. The reader's conditional verdict is right, and my stress-test identifies the same load-bearing node. The most dangerous point is not the general genericity remark itself—such arguments can sometimes be made rigorous—but that the proof of the crucial local forcing is explicitly deferred and never completed. In particular, the sentence 'We first have to make sure that the correct tiles appear around the vertex... but we'll leave that for later' occurs exactly where the theorem needs a proof, and later text relies on that step without returning to it. Similarly, the final sentence 'As now the protoset P behaves in the same way...' is a conclusion, not a proof; the recomposition definition alone does not imply equality of tiling spaces. These omissions are repairable in principle, so CONDITIONAL rather than REJECT is appropriate; however, as written the theorem is not established. I find no reason to doubt the author's good faith, and the exposition of the background is useful; the issue is purely the missing argument.","tokens_in":10150,"tokens_out":10837,"duration_ms":131070,"concrete_test":"Work out an explicit numerical or exact instance: choose α and the positions of Q,R,Z satisfying every condition listed in §3.2 (the unique 360° sum α+β+γ, no angle sum equal to θ or 180°−θ, distinct consecutive new edge lengths, convexity). Then compute the complete vertex star catalogue of the six prototiles. If no such instance exists, Theorem 3.2 is false. If an instance exists, enumerate all edge-to-edge tilings of patches of radius up to, say, 20 around a seed tile using a tiling solver or exact neighborhood expansion, and check that every patch is a sub-patch of one of Schmitt's tilings of Figure 5(b). A single non-Schmitt patch is a counterexample to the claimed forcing; reproducing only Schmitt-like patches would materially support the 'behaves in the same way' step.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 3.2 has two linked gaps. First, the required parameter choice is left to an assertion of 'almost inevitable' genericity: α and Q,R must simultaneously avoid finitely many bad angle sums, satisfy convexity, and make inner edge lengths distinct, but these are inequalities coupled through the geometry of Figure 7, so the forbidden set is not literally finite. No explicit choice or existence proof is given. Second, and more load-bearing, the proof needs to show that every vertex in every tiling by P is surrounded exactly as in Figure 7. The paper explicitly defers the key step: 'We first have to make sure that the correct tiles appear around the vertex... by a bit more complicated argument, but we'll leave that for later' (§3.2). That argument never appears. The subsequent order-forcing around Q assumes the correct tile set around the vertex has already been established. Even if a parameter choice exists, the conclusion 'P behaves in the same way that the pair of prototiles discovered by Schmitt...' is asserted, not proved: recomposition as defined in §3.2 only requires marks of a P-tiling to form some Schmitt tiling, not that the map on tilings be countable-to-one or surjective. Extra P-tilings, possibly continuum many, are not ruled out.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies σ-morphic protosets (protosets that tile the plane in exactly countably many noncongruent ways). It surveys known constructions, presents a new non-convex σ-morphic protoset of three tiles (Theorem 3.1, Figure 6), and states the main result Theorem 3.2: a protoset P consisting of convex prototiles (shown in Figure 7) that is σ-morphic, obtained by recomposing Schmitt's σ-morphic protoset into convex polygons. The proof of Theorem 3.2 is a sketch: it proposes to choose angles and edge lengths generically so that every vertex of every tile is forced to be surrounded exactly as in Figure 7, then argues that the local forcing around vertices Q and R propagates to make P behave like Schmitt's protoset. A second convex protoset (Figure 8) is also announced with its proof omitted.","tokens_in":10440,"tokens_out":9035,"duration_ms":88367,"significance":"If the proof of Theorem 3.2 can be completed, the result would be the first example of a σ-morphic protoset consisting solely of convex prototiles, answering a natural variant of the open problem of Grünbaum and Shephard and showing that convexity is not an obstruction to countability of tiling classes. The new non-convex protoset of Theorem 3.1 and the geometric ideas for subdividing bump-and-nick tiles into convex pieces would also be of independent interest. The paper is clearly written and includes a useful survey; the constructions are explicitly tied to Schmitt's prior work, and the central claim does not depend on the conclusion it sets out to prove, so there is no circularity. However, as it stands the proof of the main theorem contains substantial gaps, as detailed below.","major_comments":[{"comment":"The proof explicitly defers the key vertex-forcing step: 'We first have to make sure that the correct tiles appear around the vertex... by a bit more complicated argument, but we'll leave that for later.' This step never appears. The subsequent order-forcing arguments for vertices Q and R assume that the correct set of tiles around each vertex has already been established. Since the theorem's conclusion requires that every vertex in every tiling by P have exactly the local configuration of Figure 7, the missing argument is load-bearing: without it, extra tilings, possibly continuum many, are not excluded.","section":"§3.2, proof of Theorem 3.2"},{"comment":"The existence of a suitable angle α and of subdivision points Q, R, Z is justified by 'almost inevitable' genericity, stated as 'the set of forbidden angles is finite' and 'continuum many placements... taking one that will fulfill the requirements on the angles is then almost inevitable'. The forbidden conditions, however, include both angle equalities and edge-length equalities (for instance, distinctness of consecutive inner edges such as |QS| ≠ |QT|), which are algebraic equations and inequalities in the parameters (α and the position of QR). The bad parameter set is a finite union of proper real-algebraic subsets which is typically infinite, not a finite set. A rigorous existence proof (for example by a dimension/measure argument or an explicit construction) for a parameter choice satisfying all angle and length constraints simultaneously is needed.","section":"§3.2, genericity of the parameters α and QR"},{"comment":"The conclusion 'As now the protoset P behaves in the same way that the pair of prototiles discovered by Schmitt... we have shown that P is indeed σ-morphic' does not follow from the recomposition definition used in the paper. The definition of recomposition only requires that the marks of any P-tiling form some Schmitt tiling; it does not state that the induced map from P-tilings to Schmitt tilings is countable-to-one, nor that every Schmitt tiling admits at least one P-tiling. Even if every P-tiling marks a Schmitt tiling, there could be continuum many P-tilings marking the same Schmitt tiling. The proof must show that the fibers of this map are countable (or alternatively count P-tilings directly).","section":"§3.2, final paragraph"},{"comment":"The second convex protoset shown in Figure 8 is announced with the proof omitted ('We will, however, omit the proof as it uses similar ideas as the previous one'). Since the section presents this as an additional result, the paper either needs to provide the proof or explicitly downgrade the statement to a conjecture. As written, the claim that the protoset is σ-morphic is unproved.","section":"§3.2, Figure 8"}],"minor_comments":[{"comment":"There is a typo in the theorem statement: 'The protest shown in Figure 6 (a) is σ-morphic' should read 'protoset'.","section":"§3.1, Theorem 3.1"},{"comment":"In the definition of recomposition, 'any tiling T1 admitted by T1' should read 'any tiling admitted by T1', and similarly for 'marks on the tiles are the vertices and edges of some tiling admitted by T2' the grammar needs fixing.","section":"§3.2, recomposition definition"},{"comment":"The claim that 'each triangle is c-morphic while quadrangles can either be monomorphic or c-morphic' is stated without proof or reference; a citation or a brief justification would be helpful.","section":"§3.2, paragraph after Theorem 3.2"},{"comment":"The geometric argument concerning the edge ZW and the points Z′, Z′′, Z′′′, W′ relies on a configuration that is not fully described in the text; adding a sub-figure or a precise coordinate description would improve readability and verifiability.","section":"§3.2, proof of Theorem 3.2, tiles 7 and 8"}],"recommendation":"major_revision","confidential_remarks":"The main theorem is not established as written: the key forcing argument is deferred, the genericity claim is not rigorous, and the final countability conclusion does not follow from the recomposition definition. These are substantial but potentially fixable gaps; the approach is promising and the non-convex result of Theorem 3.1 is a nice contribution. I recommend major revision with the expectation that the missing proof be supplied or the claims be appropriately weakened."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The honest take: the paper contains a genuinely new idea and a promising construction, but its central theorem is not established as written. The author does two things. First, a new 3-tile sigma-morphic protoset with bumps and nicks (Theorem 3.1) whose proof is largely complete and is a nice addition in its own right. Second, the headline: a protoset of six convex polygons claimed to be sigma-morphic, obtained by 'convexifying' Schmitt's two-tile protoset. That would be the first convex example and a real answer to a natural question. The construction is new, and the survey of existing c-morphic and sigma-morphic machinery is useful and well-cited.\n\nThe soft spot is the proof of Theorem 3.2. The key step--forcing the correct tiles to appear around every vertex--is explicitly deferred: 'We first have to make sure that the correct tiles appear around the vertex ... but we'll leave that for later.' The later argument never appears. Everything after that, including the edge-length forcing around Q and R, assumes the surrounding tile set is already known, so the local picture is not actually proved. The genericity claim for choosing alpha and the points Q,R is plausible: the forbidden configurations form a finite union of algebraic conditions, so an open dense set of good parameters should exist. But that existence alone does not supply the missing forcing proof. Second, the step from 'P-tilings mark out Schmitt tilings' to 'P has exactly countably many tilings' is unjustified. Recomposition gives a map from P-tilings to Schmitt tilings, but one also needs to show the map is countable-to-one and that no continuum of P-tilings live over a single Schmitt tiling. That is not done. The omitted proof for the Figure 8 protoset is a minor sin in comparison.\n\nThere is no apparent fatal flaw in the construction itself; it looks repairable. But a referee should demand a complete argument before accepting the theorem. This paper deserves to be sent out for review--the problem is natural, the approach is new, and a careful referee could either fill the gaps or find a counterexample. I would not cite it as a proven result yet, but I would mention it as a construction claimed to be sigma-morphic with a proof sketch. Bring it to reading group if you want a case study in how much detail counts as proof in tiling arguments.","headline":"A promising construction of a convex sigma-morphic protoset, but the main theorem's proof leaves the crucial vertex-forcing argument deferred and the counting step unjustified, so the result is not established as written.","tokens_in":10924,"tokens_out":4737,"would_cite":false,"duration_ms":51165,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52C20","05B45","52A37"],"pacs":[],"model":"deepseek-v4-flash","headline":"A protoset of six convex polygons is constructed that tiles the plane in exactly countably many noncongruent ways.","keywords":["tiling","tessellation","σ-morphic tile","polymorphic tile","convex protoset","recomposition","convex prototiles"],"falsifier":"Take the six polygons with a proposed choice of $\\alpha$, $Q$, $R$, and $Z$; compute all sums of tile angles that equal $360^\\circ$ and all sums that equal $180^\\circ-\\theta$ for an inner angle $\\theta$. If any unintended $360^\\circ$ sum appears, or any two consecutive inner edges have equal length, the vertex surrounding is not unique and the protoset admits an extra local configuration, opening the way to uncountably many tilings.","tokens_in":9955,"feed_emoji":"🔷","tokens_out":16505,"duration_ms":169176,"temperature":0.7,"pith_summary":"This note establishes that there is a protoset $P$ consisting only of convex polygons — four hexagons and two pentagons — which is $\\sigma$-morphic: it tiles the Euclidean plane in exactly $\\aleph_0$ many noncongruent ways up to isometry. The significance is that every earlier construction of a $\\sigma$-morphic protoset used bumps and nicks, which force matching by non-convex edge shapes; the paper shows the same countably-infinite tiling behaviour can be achieved inside the class of convex prototiles. The construction is carried out by recomposing a known two-tile non-convex $\\sigma$-morphic protoset: each bump or nick is replaced by two straight edges forming a unique angle $\\alpha$, and the subdivision points are chosen so that every vertex has exactly one forced surrounding. A secondary result is a new three-tile non-convex protoset (Theorem 3.1) whose $\\aleph_0$ tilings are obtained by inserting an arbitrary finite number of red rows.","feed_headline":"Convex-only tileset achieves exactly countably many tilings","feed_subtitle":"Recomposing a known non-convex σ-morphic protoset yields convex polygons with exactly countably many tilings.","key_machinery":"The load-bearing object is the angular replacement of a bump or nick. In the base protoset, matching is forced by a protrusion entering an indentation; here each such edge is replaced by two straight edges meeting at an angle $\\alpha$, one pair protruding outward and one pair receding inward, so that only an $\\alpha$-corner of the opposite sign can fit it. The paper adds two layers of control on the subdivided polygons: the angle sums available at any vertex are restricted (the only way to reach $360^\\circ$ with $\\alpha,\\beta,\\gamma$ is $\\alpha+\\beta+\\gamma$, and no sum produces $\\theta$ or $180^\\circ-\\theta$), and the lengths of consecutive inner edges are all distinct, so a 3-valent vertex can be surrounded in only one order and orientation. Recomposition — marking the new polygons so that the original tile edges reappear — transfers the forced tilings of the base protoset to the convex protoset $P$.","core_discovery":"The paper's central claim is Theorem 3.2: there exists a protoset $P$ of convex prototiles that is $\\sigma$-morphic. In the proof, a known two-tile non-convex $\\sigma$-morphic protoset (the red outlines in Figure 7) is divided by inner segments into four convex hexagons and two convex pentagons. The subdivision is chosen so that (i) the angle $\\alpha$ of each new angular bump or nick, together with the angles $\\beta,\\gamma$, gives the only triple of tile angles summing to $360^\\circ$; (ii) no two consecutive inner edges of any tile have equal length; and (iii) no sum of tile angles equals $\\theta$ or $180^\\circ-\\theta$ for any inner angle $\\theta$. These conditions make each vertex of every tile admit exactly one possible cyclic arrangement of the tiles around it, and the edge-length inequalities force the order and orientation of the tiles. Consequently the tilings by $P$ are in bijective correspondence with the tilings by the base protoset, and since the latter has exactly $\\aleph_0$ noncongruent tilings, so does $P$.","pith_inferences":["The proof's genericity step suggests a general recipe: any finite non-convex protoset whose forcing is purely local and vertex-based could be recomposed into convex polygons by replacing each bump or nick with an angular chain and choosing the chain angle outside the finitely many forbidden sums; testing this on other base constructions would show how far the method extends.","The paper does not give explicit coordinates for $\\alpha, Q, R, Z$; a concrete instantiation would let a computer check every vertex angle sum and edge-length inequality, turning the existence theorem into a verified explicit family of convex polygons.","Because the construction uses six prototiles, a natural next test of the method is to search for a smaller convex protoset; the paper itself asks whether three or even two convex prototiles could suffice."],"forward_implications":["If the proof is correct, this is a $\\sigma$-morphic protoset whose prototiles are all convex polygons, so countably-many-tiling behaviour is compatible with convexity at the level of protosets.","The protoset $P$ inherits the tilings of the base protoset, including a periodic tiling; the paper recalls that any $\\sigma$-morphic protoset must admit a periodic tiling.","The new three-tile protoset of Theorem 3.1 gives a second mechanism for $\\aleph_0$ tilings: a discrete parameter counts the finitely many red rows, with all but two tilings obtained this way.","By the classification facts surveyed in the paper, no single convex tile can be $\\sigma$-morphic: the possible convex monohedral tilers yield either finitely many or continuum many tilings."],"supporting_citations":[{"why":"Supplies the base two-tile non-convex $\\sigma$-morphic protoset whose tilings are transferred to the convex protoset.","marker":"[24]"},{"why":"Defines recomposition of protosets and records the convex aperiodic examples that motivate the construction.","marker":"[6]"},{"why":"Provides the earlier method of forcing countably many tilings by discrete shifts, underlying the countable family here.","marker":"[21]"},{"why":"States that an aperiodic protoset has continuum many tilings, whose contrapositive gives a periodic tiling for any $\\sigma$-morphic protoset.","marker":"[4]"}],"fun_headline_variants":["Convex tiles now achieve exactly countably many tilings","σ-morphic protoset made entirely of convex pieces","Countably many tilings from convex-only tiles","Convex-only trick for aleph-null tilings"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The construction rests on the assumption that one can choose the angle $\\alpha$ and the positions of the points $Q$, $R$, and $Z$ so that all required inequalities hold simultaneously — every vertex has a unique forced surrounding, no two consecutive inner edges are equal, no forbidden angle sum appears, and every polygon stays convex — with no explicit choice actually exhibited.","fun_headline_variants_meta":{"raw":{"variants":["Convex tiles now achieve exactly countably many tilings","σ-morphic protoset made entirely of convex pieces","Countably many tilings from convex-only tiles","Convex-only trick for aleph-null tilings"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000198,"raw_usage":{"total_tokens":1349,"prompt_tokens":905,"completion_tokens":444,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":521,"completion_tokens_details":{"reasoning_tokens":380}},"tokens_in":521,"tokens_out":444,"duration_ms":5854,"temperature":1.0,"reasoning_tokens":380,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T13:11:41.659671+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the six polygons with a proposed choice of $\\alpha$, $Q$, $R$, and $Z$; compute all sums of tile angles that equal $360^\\circ$ and all sums that equal $180^\\circ-\\theta$ for an inner angle $\\theta$. If any unintended $360^\\circ$ sum appears, or any two consecutive inner edges have equal length, the vertex surrounding is not unique and the protoset admits an extra local configuration, opening the way to uncountably many tilings.","supporting_citations":[{"cited_title":"Schmitt, σ-morphic sets of prototiles, Discrete Comput","cited_arxiv_id":null,"evidence_quote":"Supplies the base two-tile non-convex $\\sigma$-morphic protoset whose tilings are transferred to the convex protoset."},{"cited_title":"Grünbaum and G","cited_arxiv_id":null,"evidence_quote":"Defines recomposition of protosets and records the convex aperiodic examples that motivate the construction."},{"cited_title":"Schmitt, Pairs of tiles which admit finitely or counta bly infinitely many tilings, Geom","cited_arxiv_id":null,"evidence_quote":"Provides the earlier method of forcing countably many tilings by discrete shifts, underlying the countable family here."},{"cited_title":"Danzer, and N","cited_arxiv_id":null,"evidence_quote":"States that an aperiodic protoset has continuum many tilings, whose contrapositive gives a periodic tiling for any $\\sigma$-morphic protoset."}],"review_version":1}