{"id":"eebfdadf-1a15-4ae4-9421-6e5824f2d849","arxiv_id":"1908.08225","paper_version":2,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"The lattice of submonoids generated by idempotents and one- or two-sided units of a monoid is classified by a four-bit type, and the functors producing these submonoids generate a 15-element monoid.","lead":"Given any algebraic object called a monoid, this paper studies the sub-pieces you can build from its idempotents and its one-sided or two-sided units. It proves those sub-pieces always form one of a small list of lattice shapes, and that the natural operators extracting them form a monoid of exactly 15 operators.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified","rationale":"The reader's ACCEPT verdict is sound. The central claims—the 15-element monoid of functors and the classification of lattice invariants by the four-bit type—are explicitly proved with elementary arguments. The only externally imported fact that could materially affect the argument is [22, Lemma 2.5], which the reader identified as the weakest assumption. I checked it directly: the identity e h = h(h' e h), valid for any h∈GR with right inverse h' and any idempotent e, shows that idempotents can be moved to the right of right units, so every word in FR reduces to a right unit followed by a product of idempotents. Thus FR=GR·E(M) holds in full generality, Lemma 3.5(ii) is valid, and the dichotomy in Lemma 3.10 together with Theorem 4.4 follows. I also checked the witness monoid M=G×E×P×B: the fifteen submonoids in L+(M) are pairwise distinct because the group, idempotent-generated, positive-integer, and bicyclic coordinates separate the functors, so |F+|=15. The unshipped GAP code is used only for secondary structural statistics in §5.5, not for the size proof or the classification. The only textual issue is a sign typo in §4 before (4.3), where the intersection of {G,GL,GR} and {F,FL,FR,FLR} should be empty rather than nonempty; the surrounding argument and Lemma 3.7(ii) make the intended correction unambiguous, so this does not undermine the theorem.","tokens_in":21866,"tokens_out":35419,"duration_ms":315633,"concrete_test":"Independently prove [22, Lemma 2.5] for an arbitrary monoid M: take any word in GR(M)∪E(M), and whenever an idempotent e precedes a right unit h, rewrite e h = h(h' e h) with hh'=1; repeat to move all right units to the left. If the resulting equality FR(M)=GR(M)E(M) holds, Lemma 3.5(ii) and Theorem 4.4 stand. Also check the printed sign in §4 before (4.3): it should read =∅, not ≠∅.","verdict_should_be":"UNCHANGED","load_bearing_attack":"After checking the one point the reader flagged, no load-bearing concern remains. The imported identity FR(M)=GR(M)E(M) from [22, Lemma 2.5] is not merely cited: it follows in one line. For h∈GR with hh'=1 and idempotent e, we have e h = h(h' e h), and h' e h is idempotent; bubbling right units left past idempotents gives every element of FR as a right unit times a product of idempotents. This validates Lemma 3.5(ii), the stable/unstable dichotomy, and the type classification. The only blemish found is in §4, before (4.3): since E≠{1}, Lemma 3.7(ii) gives {G,GL,GR}∩{F,FL,FR,FLR} = ∅, but the text prints ≠∅. The correct symbol is immediate and the distinctness claim (4.3) follows; this is a typo, not a mathematical gap.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the submonoids of an arbitrary monoid M generated by idempotents and by one- or two-sided units, viewed as functors on the category of monoids with composition as the operation; the basic functors are O, E, G, GL, GR, GLR, F, FL, FR, FLR, I. The main results are threefold. First, the lattice L(M) = {X(M) : X in F} is classified: a four-bit type T(M) records whether G = GL, whether FLR = M, whether FLR = GLR, and whether G = {1}; Proposition 4.2 shows that all sixteen types occur via direct products of a nontrivial group, an idempotent-generated monoid, the positive integers, and the bicyclic monoid, and Theorem 4.4 shows that the type completely determines L(M), whose possible shapes are the finite list in Figures 4, 6 and 7. Second, the monoid F+ generated by F together with the four new functors Q = E∘GLR, P = F∘GLR, PL = FL∘GLR, PR = FR∘GLR is proved to have exactly fifteen elements (Proposition 5.6), with a complete composition table (Table 3) and a description of its Green structure, including J-triviality and subsemigroup and congruence counts obtained with GAP (Section 5.5). Third, Section 6 classifies the enhanced lattice L+(M) = {X(M) : X in F+} and shows that it equals L(M) unless T1 = T3 = 0, in which case at most four additional nodes appear; the enhanced invariant carries no more isomorphism-type information than L.","tokens_in":22053,"tokens_out":58622,"duration_ms":486818,"significance":"If correct, the results give a complete and very small classification: every monoid's lattice of submonoids generated by idempotents and units is one of finitely many shapes determined by four yes/no questions, and the composition of any two of the fifteen functors is again one of them, so the associated operators form a 15-element J-trivial monoid. The paper's main strength is its explicitness: the composition tables are fully displayed; the distinctness claims are witnessed by four concrete monoids; all sixteen types are realized by explicit products; and the classification is falsifiable in that any monoid can be checked against its type. The paper is also honest about the limits of the invariant, including the open question of whether L+(M) is always a sublattice of Sub(M) (Remark 6.5). I explicitly checked the point most plausibly load-bearing, the identity FR(M) = GR(M)·E(M) imported from [22, Lemma 2.5] and used in Lemma 3.5(ii); it is elementary (conjugate idempotents past right units) and is not a gap. The GAP-based structural claims in Section 5.5 are stated clearly but would benefit from an accompanying script for full reproducibility.","major_comments":[],"minor_comments":[{"comment":"In the unstable case, the text says that Lemma 3.7(ii) gives {G, GL, GR} ∩ {F, FL, FR, FLR} ≠ ∅; since it has just been established that E ≠ {1}, Lemma 3.7(ii) gives the opposite, namely that this intersection is empty, and the empty intersection is exactly what is needed to justify the distinctness of the seven submonoids listed in (4.3). This looks like a typographical error in the inequality symbol rather than a mathematical gap, but it should be corrected.","section":"Section 4, before (4.3)"},{"comment":"The equality FR(M) = GR(M)·E(M) is imported from [22, Lemma 2.5] and is load-bearing, since it is used to prove that GL ∩ FR ⊆ G, which underlies the stable/unstable dichotomy (Lemma 3.10) and hence the classification in Theorem 4.4. The proof is very short (for a right unit g with right inverse g' and an idempotent e, one has eg = g(g'eg) with g'eg idempotent, so right units can be moved to the left of any product of idempotents); please include it so that the central argument is self-contained.","section":"Lemma 3.5(ii)"},{"comment":"The structural claims about F+ (J-triviality, 2904 subsemigroups, 1613 congruences, 76 principal congruences) are attributed to GAP, but no code or session output is provided, so these claims cannot be checked by a reader. Please include the GAP script or an explicit machine-readable verification artifact in an appendix or as supplementary material.","section":"Section 5.5"},{"comment":"The proof that |F+| = 15 rests on the assertion that L+(M) has size 15 for M = G×E×P×B, followed by the phrase that one may easily check this. Since this is the only place where the fifteen functors are shown pairwise distinct, please display the table of the fifteen quadruples (X(G), X(E), X(P), X(B)) or spell out which pairs of functors are separated by which factor; the preceding sentence on the lower bound |F+| ≥ 11 is also too compressed, as the point is that the eleven functors of F already give eleven distinct submonoids on a monoid of type (0,0,0,0).","section":"Proposition 5.6"},{"comment":"The symbol E is used for both the set of idempotents and the submonoid they generate, and in Lemma 3.3 and in the proof of Lemma 3.5(ii) the two uses appear close together. Please adopt a clearer convention (for instance a distinct symbol for the set of idempotents) or state the convention once in Section 2.1, since this is a genuine source of possible misreading.","section":"Sections 2.1 and 3"},{"comment":"The classification proof is a prose case analysis in which the figures carry much of the information. Please state explicitly that the displayed Hasse diagrams are verified by combining Lemmas 3.5-3.10 with the type conditions, in particular that every displayed cover is a proper containment and that no containments other than the transitive closure of the displayed edges occur; this would make the role of the figures fully checkable.","section":"Theorem 4.4 and Figures 4, 6, 7"}],"recommendation":"minor_revision","confidential_remarks":"For the editor: the manuscript is mathematically sound as far as I can verify, and I recommend acceptance after minor revision. The only required correction is the inequality symbol in Section 4; the remaining items are improvements. The dependence on the author's earlier article [22] is transparent and the imported lemmas are elementary; I see no circularity or overlap concerns. The paper fits the journal's scope well."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the headline: this paper delivers what it promises. It fully classifies the lattice L(M) of submonoids generated by idempotents and one-sided/two-sided units, showing that the whole shape is controlled by a four-bit type T(M). It also shows that the natural extraction functors generate a monoid F+ of exactly 15 elements. Both results are new, and both are proved explicitly with elementary combinatorial arguments.\n\nWhat earns credit: the proof of |F+|=15 via direct evaluation on G × E × P × B is concrete and convincing; Theorem 4.4 with Figures 4, 6, 7 gives a complete picture; and the paper is careful about when L(M) is not a sublattice of Sub(M) (Remark 2.6) and when the enhanced L+(M) may or may not be closed under meets (Remark 6.5). The author also acknowledges the open question about L+(M) being a sublattice generally, which is honest.\n\nThe soft spots are minor. The text around (4.3) has a typo: it says the intersection {G,GL,GR} ∩ {F,FL,FR,FLR} is nonempty, while Lemma 3.7(ii) gives exactly the opposite when E ≠ {1}. The correct symbol is '= ∅', and the distinctness claim in (4.3) follows anyway. Also, the GAP computations in Section 5.5 (J-triviality, 2904 subsemigroups, 1613 congruences) are stated without shipped code, so a reader cannot independently reproduce those numbers; the main structural results of the paper do not depend on them.\n\nOn circularity: the paper leans on several lemmas from the author's earlier [22], including the key identity FR(M)=GR(M)E(M). The stress-test note confirms that this identity follows in one line, and my own reading agrees, so this is not a load-bearing gap. It is a mild self-containedness issue at most.\n\nBottom line: this is a clean, well-written pure mathematics paper for semigroup theorists and anyone interested in monoid invariants. The classification is complete and the proofs are visible. I would accept it with minor revisions—fix the typo, and if possible include the GAP verification script. It deserves a serious referee.","headline":"A genuinely useful classification of idempotent/unit-generated submonoids with a 15-element functor monoid, mostly tight proofs, and only minor blemishes; worth refereeing.","tokens_in":22639,"tokens_out":2599,"would_cite":true,"duration_ms":23512,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["20M50","20M10","20M15","20M20","18D10"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the lattice of submonoids generated by idempotents and one-sided units in any monoid is determined by a four-bit type, and that the functors extracting these submonoids compose to form a monoid of exactly fifteen…","keywords":["monoids","idempotents","units","one-sided units","lattices","functors","invariants","Green's relations"],"falsifier":"Enumerate all monoids of order up to 8, compute for each monoid the type $T(M)$ and the lattice $L(M)$, and check that $L(M)$ is the diagram prescribed by Theorem 4.4; any monoid whose lattice is not the predicted shape is a direct counterexample. A more targeted test is to check whether every element of $F_R(M)$ has the form $g\\,e_1\\cdots e_k$ with $g$ a right unit and each $e_i$ idempotent, since the imported product description is the step where the argument would first fail.","tokens_in":21630,"feed_emoji":"🧩","tokens_out":16820,"duration_ms":209081,"temperature":0.7,"pith_summary":"Every monoid contains a small family of natural submonoids: subsets closed under the operation that are generated by idempotents (elements equal to their own square), by left or right units (elements with one-sided inverses), by two-sided units, and by combinations of these. This paper proves that the inclusion pattern among these submonoids, the lattice $L(M)$, is completely determined by four yes/no answers collected in a type $T(M)=(T_1(M),T_2(M),T_3(M),T_4(M))$, and that only a short finite list of lattice shapes can occur. It further shows that the operators choosing these submonoids are functors on the category of monoids and generate, under composition, a monoid of exactly fifteen functors. All sixteen possible types are realized by explicit monoids, so the classification is exhaustive rather than an artifact of missing examples. If correct, this means the collection of idempotent and unit submonoids is always one of a few rigid patterns fixed by four bits of data.","feed_headline":"Fifteen functors tie every monoid's unit lattice to four bits","feed_subtitle":"A four-bit 'type' decides which of the finite submonoid-lattice shapes a monoid has.","key_machinery":"The load-bearing mechanism is the four-bit type $T(M)$ together with the stable/unstable dichotomy from Green's relations. The identity element $1$ of a monoid is stable exactly when the $\\mathcal{J}$-class of $1$ equals the unit group $G(M)$, equivalently when $M$ contains no bicyclic submonoid; in that case $G=G_L=G_R=G_{LR}$ and the lattice collapses to the five-element chain of Figure 3. When $1$ is unstable, Lemmas 3.5 and 3.6 force the seven submonoids $G,G_L,G_R,F,F_L,F_R,F_{LR}$ to remain distinct while leaving only the three binary choices measured by $T_2,T_3,T_4$. The argument's key identity is the imported product description $F_R(M)=G_R(M)\\cdot E(M)$, which feeds into Lemma 3.5(ii) to show $G_L\\cap F_R\\subseteq G$; that containment is what prevents uncontrolled collapse and lets the type alone determine the whole lattice. The direct-product multiplicativity of the type, $T(M\\times N)=T(M)\\times T(N)$, then lets the paper realize all sixteen types from four small monoids.","core_discovery":"The paper's central claim is that the lattice $L(M)=\\{X(M):X\\in \\mathcal{F}\\}$, where $\\mathcal{F}=\\{O,E,G,G_L,G_R,G_{LR},F,F_L,F_R,F_{LR},I\\}$, is classified by the binary quadruple $T(M)=(T_1(M),T_2(M),T_3(M),T_4(M))$. The four bits ask whether $G(M)=G_L(M)$, whether $F_{LR}(M)=M$, whether $F_{LR}(M)=G_{LR}(M)$, and whether $G(M)=\\{1\\}$; Theorem 4.4 states that for a stable identity the lattice is one of the eight diagrams in Figure 4, and for an unstable identity one of the eight diagrams in Figure 6, with Figure 7 listing the possibilities up to isomorphism. The same section shows every one of the sixteen quadruples arises by taking direct products of four basic monoids. Independently, Section 5 proves that the enlarged collection $\\mathcal{F}^+=\\mathcal{F}\\cup\\{Q,P,P_L,P_R\\}$, with $Q=E\\circ G_{LR}$, $P=F\\circ G_{LR}$, $P_L=F_L\\circ G_{LR}$, $P_R=F_R\\circ G_{LR}$, is closed under composition and has exactly fifteen elements, the four new functors being genuinely new; the enhanced lattice $L^+(M)$ is then shown to be classified by the same four-bit type and hence to add no discriminating power beyond $L(M)$.","pith_inferences":["The same direct-product recipe that realizes all sixteen types also serves as a practical way to construct monoids with prescribed lattice shapes from a group, an idempotent-generated monoid, the positive integers, and the bicyclic monoid.","This suggests testing whether other natural element classes, for example regular elements or elements whose powers eventually repeat, give operators that again form a finite monoid under composition; the closure to fifteen here makes such finiteness plausible.","Remark 6.5 leaves open whether $L^+(M)$ is always a sublattice of $Sub(M)$; that can be settled by a computer search over small monoids, and a counterexample would show the enhanced lattice has a subtle incompleteness invisible to the four-bit classification.","The four added functors are best read as bookkeeping that closes the operator monoid under composition rather than as new invariants, since they change neither the type nor the discriminating power of the lattice."],"forward_implications":["For any monoid $M$, the lattice $L(M)$ has one of the finitely many shapes in Figure 7; deciding which one requires only the four bits of $T(M)$.","Every one of the sixteen types occurs, so the classification cannot be sharpened by adding further binary conditions of the same kind to the type.","The eleven functors of $\\mathcal{F}$ together with $Q,P,P_L,P_R$ form a 15-element monoid under composition; no further new functors appear when all compositions are taken.","The enhanced lattice $L^+(M)$ is classified by the same four-bit type as $L(M)$, so it adds no new discriminating information beyond $L(M)$.","The monoid $\\mathcal{F}^+$ is $\\mathcal{J}$-trivial, so its Green's relations coincide with equality and the divisibility order in Figure 9 describes the whole structure."],"supporting_citations":[{"why":"Supplies the imported lemmas about idempotents and one-sided units, including the product description for the submonoid generated by right units and idempotents used in Lemma 3.5(ii), and the basic identities behind the composition table.","marker":"[22]"},{"why":"Provides the bicyclic monoid, Green's relations background, and basic monoid facts on which the stability dichotomy and several examples rely.","marker":"[45]"},{"why":"Supplies the standard terminology for idempotents and left, right, and two-sided units used throughout the paper.","marker":"[7]"},{"why":"Introduces Green's relations, which underpin Lemma 3.10 and the stable/unstable split of the classification.","marker":"[39]"},{"why":"The software package used to verify the structural claims about the 15-element monoid, including J-triviality and the divisibility diagram.","marker":"[58]"},{"why":"Supplies the monoidal category background used to view the operators as monoidal functors and to prove the direct-product multiplicativity of the type.","marker":"[51]"}],"fun_headline_variants":["15 functors, 4 bits: all monoid lattices classified","Fifteen functors, four bits: a complete lattice classification for monoids","A 15-element functor monoid reveals all monoid sublattice shapes via 4 bits","Idempotents and one-sided units: 15 functors classify every monoid lattice","Four-bit code decides a monoid's lattice; functor monoid has 15 members"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything rests on the imported lemma that in every monoid the submonoid generated by right units and idempotents consists exactly of products of one right unit with finitely many idempotents; if some monoid failed that description, the proof that left units meet it only in the two-sided units would break, and with it the stable/unstable collapse analysis and the classification.","fun_headline_variants_meta":{"raw":{"variants":["15 functors, 4 bits: all monoid lattices classified","Fifteen functors, four bits: a complete lattice classification for monoids","A 15-element functor monoid reveals all monoid sublattice shapes via 4 bits","Idempotents and one-sided units: 15 functors classify every monoid lattice","Four-bit code decides a monoid's lattice; functor monoid has 15 members"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001435,"raw_usage":{"total_tokens":5847,"prompt_tokens":1070,"completion_tokens":4777,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":686,"completion_tokens_details":{"reasoning_tokens":4667}},"tokens_in":686,"tokens_out":4777,"duration_ms":504996,"temperature":1.0,"reasoning_tokens":4667,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:45:52.242631+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Enumerate all monoids of order up to 8, compute for each monoid the type $T(M)$ and the lattice $L(M)$, and check that $L(M)$ is the diagram prescribed by Theorem 4.4; any monoid whose lattice is not the predicted shape is a direct counterexample. A more targeted test is to check whether every element of $F_R(M)$ has the form $g\\,e_1\\cdots e_k$ with $g$ a right unit and each $e_i$ idempotent, since the imported product description is the step where the argument would first fail.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the imported lemmas about idempotents and one-sided units, including the product description for the submonoid generated by right units and idempotents used in Lemma 3.5(ii), and the basic identities behind the composition table."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the bicyclic monoid, Green's relations background, and basic monoid facts on which the stability dichotomy and several examples rely."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the standard terminology for idempotents and left, right, and two-sided units used throughout the paper."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces Green's relations, which underpin Lemma 3.10 and the stable/unstable split of the classification."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The software package used to verify the structural claims about the 15-element monoid, including J-triviality and the divisibility diagram."}],"review_version":1}