{"id":"e2d4097d-206d-46cb-8021-976147fbfa70","arxiv_id":"2508.19360","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A convergent rewriting system for the Temperley-Lieb algebra is exhibited whose normal forms are the classical Jones normal forms, but the oriented version's basis claim is left as a conjecture.","lead":"This math internship report searches for a basis of the Temperley-Lieb algebra, a structure from statistical physics and knot theory, by building rewriting rules that simplify algebraic words and diagrams. It finds convergent rewriting rules for the ordinary algebra whose normal forms match the known Jones basis, but the basis claim for the oriented version rests on an unproven conjecture.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Oriented algebra basis rests on unproved category-algebra identification (Remark 4.14); without it, Theorem 4.15 gives only Hom-space bases, not a TLO_{n,k}(q) basis.","rationale":"The Reader’s verdict (CONDITIONAL) identifies the category-algebra identification as the weakest assumption, and my reading confirms this is the single most load-bearing gap. The paper’s strongest advertised claim—that rewriting yields a basis of the oriented Temperley-Lieb algebra—requires Remark 4.14, which the author explicitly states is suggested but not proved. The proof of Theorem 4.15 is also skeletal (critical pairs summarized in a figure, confluence modulo asserted via [Mal19]), but even if Theorem 4.15 is accepted, the final step to an algebra basis is missing. The non-oriented part of the paper is in better shape: the rewriting system of §2.2.2 is plausibly convergent and the comparison with Jones normal forms is given by an algorithm, though some confluence cases are summarized rather than fully displayed. The report is candid about its limitations, and the gaps are addressable rather than fatal. Since the Reader already reached CONDITIONAL on essentially these grounds, my stress-test does not change the verdict.","tokens_in":18989,"tokens_out":2962,"duration_ms":35973,"concrete_test":"Prove or disprove Remark 4.14 by a direct comparison. Concretely: fix small parameters, e.g. n=2, k=1 and n=3, k=1, enumerate the normal forms of the convergent system in §4.4 between all boundary words w,v containing k occurrences of ∨, and compare the resulting dimensions of Hom(w,v) with the known dimensions of the weight spaces of TLO_{n,k}(q) from [Bow+24]. The dimensions must match for every pair (w,v), and one must then define composition/tensor operations on these Hom-spaces that reproduce the presentation of Definition 3.3. If the dimensions match but the algebra structure cannot be recovered, Remark 4.14 is false; if they do not match, the conjectured isomorphism fails immediately. A positive result would require a rigorous proof that the categorical Hom-spaces and the algebra are isomorphic, not merely dimension counts.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper’s advertised goal is a rewriting-based basis for the oriented Temperley-Lieb algebra. Theorem 4.15 establishes convergence of a rewriting system on the oriented Temperley-Lieb category TLO(q), and §4.4 concludes that normal forms form a basis of each Hom(v,w). To convert this into a basis of the oriented Temperley-Lieb algebra TLO_{n,k}(q), one must identify the Hom-spaces between boundary words with k occurrences of ∨ with the algebra. That identification is exactly Remark 4.14, which is explicitly conjectural: 'everything suggests that the set of morphisms... is isomorphic to TLO_{n,k}(q)'. No proof is supplied. Proposition 4.12, the analogous non-oriented identification, is also only cited from [Abr09] rather than proved. The presented-category construction itself is admitted to be incomplete: in §4.2 the author writes that the proofs are not given, that the construction may lack completeness, and Remark 4.8 says the bifunctor property 'should be shown'. Thus the central oriented claim is not established: the normal forms give bases of Hom(v,w) inside the category, but whether these assemble into a basis of the oriented Temperley-Lieb algebra depends on an unproved isomorphism. This is a missing load-bearing link, not a demonstrated contradiction: the non-oriented program is plausibly sound, and the oriented program is honestly flagged as incomplete.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper, an internship report in mathematics, pursues the Temperley-Lieb algebra and its oriented analogue from the perspective of rewriting systems. In the first part, it defines TL algebra both diagrammatically and by generators and relations, recounts the classical Jones normal form, and then proposes a convergent rewriting system for the monoid generating TL_n(δ) (Theorem 2.18), with rules (1)–(6), whose normal forms it claims coincide with Jones normal forms (Theorem 2.19, via an algorithm). This yields a basis for TL_n(δ). In the second part, the paper introduces the oriented Temperley-Lieb algebra TLO_{n,k}(q), then moves to a presented strict monoidal category, the Temperley-Lieb category TL(δ) and the oriented category TLO(q). Theorem 4.15 claims that a rewriting system on TLO(q), with rules (1-left/right), (2-left/right), (3), (4), is convergent modulo the exchange relation. The paper concludes that the normal forms of this system give a basis of each Hom-space, and, via Remark 4.14, that this gives a basis of the oriented Temperley-Lieb algebra.","tokens_in":19230,"tokens_out":1785,"duration_ms":21208,"significance":"The non-oriented part demonstrates an original algorithmic and rewriting-based route to a known basis (Jones normal form) of the Temperley-Lieb algebra. The oriented part aims to do the same for the oriented Temperley-Lieb algebra, a subject with recent activity (e.g., Bowman et al.). The explicit convergence statement for the oriented category rewriting system (Theorem 4.15) is a potentially useful and clean result, and the visual rewriting rules in the category setting are promising. However, the paper's strongest advertised claim—that this yields a basis of the oriented Temperley-Lieb algebra TLO_{n,k}(q)—is not established: it rests on an identification (Remark 4.14) that the author explicitly leaves as a conjecture, and on a category construction whose completeness is acknowledged as unproved (§4.2, Remark 4.8). The manuscript is also candid about these gaps, which is a virtue, but the central theorem as stated in the abstract is therefore stronger than what is proved. If the conjectural identifications are supplied, the approach would provide a genuinely useful basis theorem for TLO_{n,k}(q).","major_comments":[{"comment":"The conclusion that normal forms of the convergent rewriting system of Theorem 4.15 form a basis of the oriented Temperley-Lieb algebra TLO_{n,k}(q) requires the identification Hom_{TLO(q)}(v,w) ≅ TLO_{n,k}(q) for v,w ∈ L_k. This identification is stated in Remark 4.14 only as 'everything suggests', with no proof. Without it, Theorem 4.15 yields bases of Hom-spaces in the presented category, not a basis of the algebra. This is a load-bearing missing link for the paper's central claim.","section":"§4.4 and Remark 4.14"},{"comment":"The presented strict monoidal category construction underlying all category-theoretic conclusions is admittedly incomplete. The author states in §4.2 that the proofs 'haven't actually' been done, that the construction 'may lack completeness', and Remark 4.8 says the bifunctor property 'should be shown'. Since the confluence proof of Theorem 4.15 is formulated inside this construction, a rigorous proof that the construction is indeed a well-defined strict monoidal linear category is needed before the Hom-space basis claim is fully justified.","section":"§4.2, Definition 4.7 and Remark 4.8"},{"comment":"The non-oriented identification End_{TL(δ)}(n) ≅ TL_n(δ) is discharged entirely by a citation to [Abr09] with no detailed argument. This is acceptable as background if the cited reference is standard, but the paper then transfers this identification to the oriented setting (Remark 4.14) without an analogous cited proof. At minimum, the author should state precisely which theorem in [Abr09] gives the endomorphism-algebra isomorphism, and explain how the presentation used here (with exchange relation) matches the standard Temperley-Lieb category.","section":"§4.3, Proposition 4.12"},{"comment":"The paper motivates the categorical approach by abandoning the direct word-rewriting system for TLO_{n,k}(q), but it does not clearly state the relation between the category TLO(q) and the algebra TLO_{n,k}(q). The final sentence of §4.4 says normal forms obtained in the category are not Jones normal forms for the non-oriented case, which is fine, but the reader is left without a precise functor or isomorphism linking Hom-sets of TLO(q) to the algebra. This gap is directly tied to the missing proof of Remark 4.14.","section":"§3.2 and §4.4"}],"minor_comments":[{"comment":"The abstract claims rewriting 'easily obtain[s] a basis' for the oriented algebra; given the conjectural status of Remark 4.14, the abstract overstates the proved content. Suggest rephrasing to indicate that the basis is obtained for Hom-spaces of the oriented Temperley-Lieb category.","section":"Abstract"},{"comment":"The proof of local confluence is a hand-check of critical pairs described in figures. The argument is plausible, but the figures for the added rules (5) and (6) are not fully detailed for all index ranges; in particular, the statement of rule (5) uses indices k ∈ [2,n−2], while the preceding discussion uses k ≤ n−3. Clarify the index bounds.","section":"§2.2.2, Theorem 2.18"},{"comment":"The correctness proof of the algorithm is written in a conversational style with several 'we can therefore consider' steps. For a formal proof, the recursive calls on subwords must be shown to terminate on words with fewer generators or lower lexicographic order; currently the measure is only stated informally. Also, the notation v′_1, v″_1 is used before being defined in the pseudocode.","section":"§2.2.3, Algorithm 2.2"},{"comment":"The oriented algebra is defined over Z[q,q−1] with generators 1_λ and e_i, but the rewriting table in §3.2 uses q as a letter and introduces 1_λ e_i 1_μ elements. The transition from the algebra presentation to the monoid-like words is not fully formal; in particular, the idempotent relations 1_λ e_i 1_λ → 0 and 1_λ 1_μ → δ_{λ,μ} 1_λ are listed as rewriting rules but their termination is not discussed.","section":"§3.1, Definition 3.3"},{"comment":"The confluence proof treats only one critical pair, saying the other is orientation-symmetric. While plausible, the proof would be stronger if it explicitly listed both critical pairs and their confluence diagrams, especially because the rules (1-left/right) and (2-left/right) are drawn in Appendix A without labels that match the theorem's numbering.","section":"§4.4, Theorem 4.15"},{"comment":"The reference [RS14] is cited for the Jones normal form and the standard modules, but the precise statements used (Theorem 2.4 and Proposition 2.6) are not attributed to specific locations in that paper. Also, [Mal19] is cited for Newman's lemma and modulo rewriting, but no page or chapter is given.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is an internship report, and its value is partly pedagogical: it gives a concrete rewriting derivation of the Jones normal form basis and a clean convergence statement for an oriented category rewriting system. The main obstacle to acceptance is the gap between the title/abstract (basis of the oriented Temperley-Lieb algebra) and the proved statements (basis of Hom-spaces in a category whose isomorphism to the algebra is conjectural). The author is honest about this gap, but the manuscript as submitted does not deliver the advertised result. I recommend major revision: either prove the categorical identifications, or reframe the paper's claims to match what is actually proved. I would not recommend rejection, as the non-oriented rewriting work is sound and the oriented categorical rewriting system is a credible starting point for future proofs."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The non-oriented half of this report is real work: Theorem 2.18 gives an explicit convergent rewriting system for the presented Temperley–Lieb algebra, and Theorem 2.19 gives a rewriting-only algorithm that produces Jones normal forms. That is a new and checkable contribution, and the critical-pair analysis, though summarized in places, is the right kind of evidence. The termination argument is straightforward, and the added rule families (5) and (6) are exactly what you need to repair the non-confluent overlap. I believe the non-oriented claims are sound, modulo a fully written confluence proof. The soft spot is the oriented program, and it is load-bearing. The abstract says the category-theoretic rewriting lets you 'easily obtain a basis for the algebra.' But the algebra basis never actually appears. Theorem 4.15 proves convergence for the oriented category TLO(q), so you get bases for each Hom(v,w) inside that presented category. To get a basis of the oriented Temperley–Lieb algebra TLO_{n,k}(q), you need to identify those Hom-spaces with the algebra. That identification is exactly Remark 4.14, which is explicitly conjectural: 'everything suggests.' Proposition 4.12, the non-oriented identification, is cited from [Abr09] rather than proved, and the author admits in §4.2 that the presented-category construction was never proved complete and that the bifunctor property 'should be shown.' So the final conclusion overreaches. The stress-test note is right: the missing category-algebra identification is a gap, not a contradiction. I want to be fair: the author flags all of this honestly. This is a candid internship report, not a polished paper. The non-oriented results deserve to be recorded, and the oriented part is a reasonable conjecture with a plausible route. The confluence checks are hand-verified and 'similar cases' are skipped; that makes independent verification harder, but not impossible. A machine-checked critical-pair verification would settle it. Who is this for? Someone working on rewriting theory for diagrammatic algebras, or on presentations of Temperley–Lieb–like categories. It is not a must-read for mainstream representation theory, but it is a serious contribution. I would send it to peer review: the non-oriented result is solid enough to merit referee time, and the reviewer can fairly ask the author to either prove Remark 4.14 or rewrite the abstract to match the actual theorems.","headline":"A genuinely useful rewriting-system result for TL_n(δ), but the oriented algebra basis advertised in the abstract rests on a conjecture the paper explicitly leaves open.","tokens_in":717,"tokens_out":821,"would_cite":false,"duration_ms":26902,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["16S15","18M05","68Q42"],"pacs":[],"model":"deepseek-v4-flash","headline":"Rewriting rules terminate at the Jones normal forms for the Temperley-Lieb algebra, and the same strategy, run in a presented category, produces bases for the oriented analogue.","keywords":["Temperley-Lieb algebra","rewriting system","Jones normal form","basis construction","oriented Temperley-Lieb algebra","Temperley-Lieb category","convergent rewriting","presented monoidal category"],"falsifier":"Take n=4 or 5 and exhaustively apply the rules of Theorem 2.18 to all words up to a bounded length; if any word has two different reduction paths ending in two different irreducible words, the system is not confluent and the normal forms do not give a unique basis. For the oriented category, spelling out End(2) in TL_O(q) explicitly and checking whether it is isomorphic to the corresponding oriented algebra would test the identification cited in Proposition 4.12 and conjectured in Remark 4.14.","tokens_in":18759,"feed_emoji":"🧮","tokens_out":10711,"duration_ms":107252,"temperature":0.7,"pith_summary":"The paper's project is to find bases for Temperley-Lieb algebras algorithmically, by setting up directed replacement rules and proving they always stop at a unique irreducible word. For the classical Temperley-Lieb algebra TL_n(δ), it exhibits six families of rewrite rules, proves the system convergent, and then shows by an explicit algorithm that the unique normal forms are exactly the Jones normal forms—so the normal forms are a basis. For the oriented Temperley-Lieb algebra, it shifts the setting to a presented strict monoidal category with caps, cups, and an exchange relation; there the same rewriting strategy is again convergent, so the normal forms of morphisms form a basis of each Hom(v,w). This matters because it replaces a basis construction usually done by combinatorial enumeration with a general mechanism: once convergence is proved, unique normal forms exist automatically and a finite rule list computes them.","feed_headline":"Rewrite rules land on the Jones normal-form basis","feed_subtitle":"Convergent rules leave every word one irreducible form; a category version extends this to oriented Temperley-Lieb.","key_machinery":"The machinery is a rewriting system: a set of words (or morphisms) equipped with directed rules that replace one subword by another. The paper proves convergence by checking termination, via a lexicographic order on words or a decreasing count of generators, and confluence, by examining critical pairs where rules overlap; a standard lemma then upgrades local confluence plus termination to full confluence. In the unoriented case, the critical-pair analysis forces two extra rules (5) and (6) that make the system confluent, and the normal forms coincide with Jones normal forms, i.e. products of antidiagonal blocks (e_i e_{i-1} ... e_j). In the oriented-category case, the key additional datum is","core_discovery":"The central claim is that basis questions for Temperley-Lieb algebras can be settled by proving a rewriting system convergent. In the unoriented case (Theorem 2.18), the rules move δ past e_i, send e_i^2 to δe_i, collapse e_i e_{i±1} e_i to e_i, swap distant generators, and add two longer collapsing rules needed after completion; the system terminates and is locally confluent, hence convergent. Theorem 2.19 then gives a rewriting algorithm whose output is always in Jones normal form—a product of antidiagonal words (e_i e_{i-1} ... e_j)—so the unique normal forms are precisely the Jones normal forms and therefore a basis of TL_n(δ). In the oriented case (Theorem 4.15), the author presents the","pith_inferences":["The same 'presented category plus rewriting' template should apply to other diagrammatic algebras (braid, Hecke, BMW) whenever they admit a finite monoidal presentation; the main obstacle would be finding the right exchange relation or modulo rules.","A natural next experiment is to fill the gap in Remark 4.14: prove that morphisms between words with exactly k occurrences of ∨ form the oriented algebra; if true, the category basis becomes an algebra basis and the oriented analogue of Theorem 2.8 follows by the same argument.","The author's warning that the free presented-category construction was not fully checked suggests the most fragile point is not confluence but the completeness of the presentation; a rigorous proof of the freeness or completeness of Definition 4.7 would harden the whole approach.","Since the category normal forms differ from Jones normal forms under the natural embedding, the oriented case likely has its own normal-form combinatorics, which could be mined for a direct combinatorial description of the basis."],"forward_implications":["Every element of TL_n(δ) has a unique normal form, computed by the six rule families; the normal forms are exactly the Jones normal forms, so they give an explicit basis without enumerating diagrams or paths.","The same rewriting argument gives a constructive proof that the presented algebra and the diagrammatic algebra are isomorphic: the map e_i ↦ E_i sends a basis to a basis.","For the oriented Temperley-Lieb category TL_O(q), each morphism space Hom(v,w) has a finite basis of irreducible morphisms, and this basis is obtained by a terminating rule set rather than by combinatorial counting.","If the conjectural identification of the category's morphism spaces with the oriented Temperley-Lieb algebra is proved, the category rewriting system yields an algebra basis for TLOn,k(q) automatically.","The convergence proof also produces a decision procedure for equality in the algebra: two words are equal exactly when their normal forms agree."],"supporting_citations":[{"why":"Supplies the bridge bijection, the increasing-path counting, and the Jones normal form method the paper uses to build its basis and to prove Theorem 1.4.","marker":"[RS14]"},{"why":"Supplies the rewriting-theory results (termination, confluence, local confluence criterion, modulo rewriting) used to prove Theorems 2.18 and 4.15.","marker":"[Mal19]"},{"why":"States the isomorphism End_TL(δ)(n) ≅ TL_n(δ) that Proposition 4.12 cites, connecting the category rewriting result to the algebra basis.","marker":"[Abr09]"},{"why":"Provides the presentation, relations, and known basis context for the oriented Temperley-Lieb algebra used in Section 3 and Remark 4.14.","marker":"[Bow+24]"},{"why":"Supplies the category-theoretic definitions of strict monoidal categories and bifunctors on which the presented-category construction of Section 4 relies.","marker":"[Mac71]"},{"why":"Supplies the free-monoid and presented-monoid formalism adapted in Definitions 1.6-1.7 and the presented-category construction.","marker":"[Deh19]"}],"fun_headline_variants":["Rewriting systems crack Temperley-Lieb basis search","Convergent rules yield Jones normal-form basis","Oriented TL basis via category theory rewriting","Algorithmic basis for Temperley-Lieb algebras","Rewrite rules pin down TL basis in both versions"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The step that carries the whole conclusion is the claim that self-maps of n in the Temperley-Lieb category match the Temperley-Lieb algebra (cited from [Abr09] without proof here), plus the conjectural matching of the oriented category's morphism spaces with the oriented algebra; if those identifications fail, the normal forms still give bases of morphism spaces but not of the algebras.","fun_headline_variants_meta":{"raw":{"variants":["Rewriting systems crack Temperley-Lieb basis search","Convergent rules yield Jones normal-form basis","Oriented TL basis via category theory rewriting","Algorithmic basis for Temperley-Lieb algebras","Rewrite rules pin down TL basis in both versions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000187,"raw_usage":{"total_tokens":1103,"prompt_tokens":618,"completion_tokens":485,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":362,"completion_tokens_details":{"reasoning_tokens":414}},"tokens_in":362,"tokens_out":485,"duration_ms":5652,"temperature":1.0,"reasoning_tokens":414,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T15:52:48.272258+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take n=4 or 5 and exhaustively apply the rules of Theorem 2.18 to all words up to a bounded length; if any word has two different reduction paths ending in two different irreducible words, the system is not confluent and the normal forms do not give a unique basis. For the oriented category, spelling out End(2) in TL_O(q) explicitly and checking whether it is isomorphic to the corresponding oriented algebra would test the identification cited in Proposition 4.12 and conjectured in Remark 4.14.","supporting_citations":[],"review_version":1}