OctoInformBeta

OpenAI preprint, September 2026 · §2. Integral Lie algebras and the special identity

Proposition 2.5 [Saturated reduction]: The Drinfeld equations survive coefficient change, with integral lifts for solution subspaces in weights greater than two

Saturated coordinate lattices retain dimension after reduction.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

Write LS=Lie⁡S⟨x,y⟩L_S=\operatorname{Lie}_S\langle x,y\rangle for the ordinary free Lie algebra over a commutative ring SS, and put L=LℚL=L_{\mathbb Q}. For a Lie polynomial qq over SS and elements a,ba,b of an SS-Lie algebra, the substitution q(a,b)q(a,b) is the image under the unique Lie homomorphism sending xx to aa and yy to bb.✦ The letter xx is the first free generator of LSL_S. The letter yy is the second free generator of LSL_S.

For a strand set II, the infinitesimal braid algebra 𝔱S(I)\mathfrak t_S(I) is the quotient of the free Lie algebra on ordered pairs (i,j)(i,j) by the Lie ideal generated by diagonal generators, differences of reversed generators, brackets with disjoint endpoint pairs, and the three-strand relations.✦ Its chord generator tijt_{ij} is the image of the free generator cijc_{ij}, and these relations are cii=0,cij−cji=0,[cij,ckl]=0({i,j}∩{k,l}=⌀),[cij,cik+cjk]=0(i,j,k distinct).\begin{gathered} c_{ii}=0,\qquad c_{ij}-c_{ji}=0,\qquad [c_{ij},c_{kl}]=0\quad\bigl(\{i,j\}\cap\{k,l\}=\varnothing\bigr),\\ [c_{ij},c_{ik}+c_{jk}]=0\quad(i,j,k\text{ distinct}). \end{gathered}(1) For rational solutions, 𝔱4\mathfrak t_4 denotes the quotient of the rational free chord Lie algebra on the unordered pairs in {0,1,2,3}\{0,1,2,3\} by the last two families of (1); here cij=cjic_{ij}=c_{ji} for distinct indices and cii=0c_{ii}=0, and tijt_{ij} again denotes the quotient image.

The antisymmetry, three-term and special defects aSa_S, hSh_S and sSs_S are the linear operators on LSL_S defined by aS(q)=q(x,y)+q(y,x),hS(q)=q(x,y)+q(y,−x−y)+q(−x−y,x),sS(q)=[x,q(y,x)]+[−x−y,q(y,−x−y)].\begin{aligned} a_S(q)&=q(x,y)+q(y,x),\\ h_S(q)&=q(x,y)+q(y,-x-y)+q(-x-y,x),\\ s_S(q)&=[x,q(y,x)]+[-x-y,q(y,-x-y)]. \end{aligned}(2) The pentagon operator ΠS\Pi_S is the linear map from LSL_S to 𝔱S({0,1,2,3})\mathfrak t_S(\{0,1,2,3\}) defined by ΠS(q)=q(t01,t12+t13)+q(t02+t12,t23)−q(t12,t23)−q(t01+t02,t13+t23)−q(t01,t12).\begin{aligned} \Pi_S(q)={}&q(t_{01},t_{12}+t_{13})+q(t_{02}+t_{12},t_{23})\\ &-q(t_{12},t_{23})-q(t_{01}+t_{02},t_{13}+t_{23})-q(t_{01},t_{12}). \end{aligned}(3) Write Πrat\Pi^{\mathrm{rat}} for the expression (3) evaluated in 𝔱4\mathfrak t_4. The Drinfeld solution submodule 𝒟S\mathcal D_S of LSL_S is the submodule on which aSa_S, hSh_S and ΠS\Pi_S vanish; these vanishing conditions are the Drinfeld equations over SS. A rational Drinfeld solution q∈Lq\in L is a Lie polynomial with vanishing defects aℚa_{\mathbb Q}, hℚh_{\mathbb Q} and Πrat\Pi^{\mathrm{rat}}; the rational solution space W⊂LW\subset L is the subspace of all rational Drinfeld solutions.

Homogeneous Coordinates and Lattices

Let α\alpha be an alphabet with nonnegative integer letter weights wt\mathrm{wt}. A bracketed word ww on the alphabet α\alpha is a letter of α\alpha or a bracket of two bracketed words on α\alpha; its evaluation replaces letters by free generators and brackets by ordinary Lie brackets. The underlying word of a bracketed word on α\alpha is the sequence of its letters read from left to right; the weight of a word on α\alpha is the sum of its letter weights. A Lie word q∈Lie⁡S⟨α⟩q\in\operatorname{Lie}_S\langle\alpha\rangle of weight nn is the evaluation of a bracketed word on α\alpha with total weight nn; the weight-nn Lie component Hwt,n(S)H_{\mathrm{wt},n}(S) is the SS-submodule of the free Lie algebra spanned by all Lie words of weight nn. In particular, for unit weights on x,yx,y, write Hn(S)H_n(S) for this component, Ln=Hn(ℚ)L_n=H_n(\mathbb Q) and Wn=W∩LnW_n=W\cap L_n.

The free associative algebra AS(α)A_S(\alpha) is the monoid algebra of words on α\alpha.✦ A word denotes its monomial with coefficient one, including the one-letter monomial associated with each associative generator. For a set PP of words on α\alpha, the prescribed word span of PP is the submodule of AS(α)A_S(\alpha) spanned by the monomials indexed by PP, and the word coefficient map for PP is the linear map from this span to SPS^P that extracts the coefficients indexed by PP. The associative component Awt,n(S)A_{\mathrm{wt},n}(S) is the span for the set Ωwt,n\Omega_{\mathrm{wt},n} of words of weight nn.✦ The universal associative embedding is the Lie homomorphism sending free Lie generators to associative generators; its restriction gives the homogeneous associative embedding of Hwt,n(S)H_{\mathrm{wt},n}(S) into Awt,n(S)A_{\mathrm{wt},n}(S).

The Lie word coordinate map cS,wt,nc_{S,\mathrm{wt},n} is the linear map on Hwt,n(S)H_{\mathrm{wt},n}(S) that records the coefficients of the homogeneous associative expansion.✦ With unit weights, abbreviate its index set to Ωn\Omega_n and its map to cS,nc_{S,n}, and use the same notation for its restriction to homogeneous solution spaces. The coordinate map is cS,wt,n:Hwt,n(S)→SΩwt,n,cS,wt,n(q)w=[w]q,c_{S,\mathrm{wt},n}\colon H_{\mathrm{wt},n}(S) \longrightarrow S^{\Omega_{\mathrm{wt},n}},\qquad c_{S,\mathrm{wt},n}(q)_w=[w]q,(4) where [w]q[w]q denotes the coefficient of ww in the associative expansion.

Integral coefficient change on an alphabet α\alpha is the Lie homomorphism βS,α:Lie⁡ℤ⟨α⟩→Lie⁡S⟨α⟩\beta_{S,\alpha}\colon\operatorname{Lie}_{\mathbb Z}\langle\alpha\rangle\to \operatorname{Lie}_S\langle\alpha\rangle fixing the free generators; its restriction to Hwt,n(ℤ)H_{\mathrm{wt},n}(\mathbb Z) is homogeneous Lie base change. For the alphabet {x,y}\{x,y\}, abbreviate this map to βS\beta_S.

For a finite set ι\iota and a rational subspace U⊂ℚιU\subset\mathbb Q^\iota, the integral coordinate lattice Λ(U)\Lambda(U) is the integer submodule of ℤι\mathbb Z^\iota consisting of families whose images under the coordinate inclusion j:ℤι→ℚιj\colon\mathbb Z^\iota\to\mathbb Q^\iota lie in UU.✦ For a field FF of arbitrary characteristic, the scalar extension map σF,U\sigma_{F,U} of the integral coordinate lattice Λ(U)\Lambda(U) is the linear map in (5), and the coordinate reduction red⁡F(U)\operatorname{red}_F(U) is its image: j(u)i=ui,Λ(U)=j−1(U),σF,U:F⊗ℤΛ(U)→Fι,σF,U(a⊗u)i=a(ui⋅1F),red⁡F(U)=im⁡σF,U.\begin{gathered} j(u)_i=u_i,\qquad \Lambda(U)=j^{-1}(U),\\ \sigma_{F,U}\colon F\otimes_{\mathbb Z}\Lambda(U)\longrightarrow F^\iota, \qquad \sigma_{F,U}(a\otimes u)_i=a\,(u_i\cdot1_F),\\ \operatorname{red}_F(U)=\operatorname{im}\sigma_{F,U}. \end{gathered}(5) Over 𝔽2\mathbb F_2, the reduced degree-nn solution space 𝒮n\mathcal S_n is the subspace (6), defined by the homogeneous condition and the antisymmetry, special and pentagon operators of (2) and (3): 𝒮n=Hn(𝔽2)∩ker⁡a𝔽2∩ker⁡s𝔽2∩ker⁡Π𝔽2.\mathcal S_n=H_n(\mathbb F_2)\cap\ker a_{\mathbb F_2} \cap\ker s_{\mathbb F_2}\cap\ker\Pi_{\mathbb F_2}.(6)✦ For a rational subspace E⊂WnE\subset W_n, the coordinate image UEU_E is the subspace cℚ,n(E)c_{\mathbb Q,n}(E) of rational word coordinates. The solution lattice reduction ℛn(E)\mathcal R_n(E) is the subspace of 𝒮n\mathcal S_n whose reduced word coordinates belong to red⁡𝔽2(UE)\operatorname{red}_{\mathbb F_2}(U_E).✦ The solution coordinate reduction ρE\rho_E is the linear map from ℛn(E)\mathcal R_n(E) to red⁡𝔽2(UE)\operatorname{red}_{\mathbb F_2}(U_E) that records reduced word coordinates, as in (7): ℛn(E)={p∈𝒮n:c𝔽2,n(p)∈red⁡𝔽2(UE)},ρE:ℛn(E)→red⁡𝔽2(UE),ρE(p)=c𝔽2,n(p).\begin{gathered} \mathcal R_n(E)=\{p\in\mathcal S_n: c_{\mathbb F_2,n}(p)\in\operatorname{red}_{\mathbb F_2}(U_E)\},\\ \rho_E\colon\mathcal R_n(E)\longrightarrow \operatorname{red}_{\mathbb F_2}(U_E),\qquad \rho_E(p)=c_{\mathbb F_2,n}(p). \end{gathered}(7)

The Claim

Theorem 1 (Saturated reduction). Let ι\iota be finite, let U⊂ℚιU\subset\mathbb Q^\iota be a rational linear subspace, and let FF be a field. For each nonzero integer kk, saturation takes the form (8): {u∈ℤι:ku∈Λ(U)}=Λ(U).\{u\in\mathbb Z^\iota:ku\in\Lambda(U)\}=\Lambda(U).(8)✦ The scalar map has kernel (9), and its image has dimension (10): ker⁡σF,U=0.\ker\sigma_{F,U}=0.(9)✦ dim⁡Fred⁡F(U)=dim⁡ℚU.\dim_F\operatorname{red}_F(U)=\dim_{\mathbb Q}U.(10)✦

Let nn be a nonnegative integer and E⊂WnE\subset W_n a rational linear subspace. The solution coordinate map has kernel (11): ker⁡ρE=0.\ker\rho_E=0.(11)✦ For n>2n>2, the dimension is (12): dim⁡𝔽2ℛn(E)=dim⁡ℚE.\dim_{\mathbb F_2}\mathcal R_n(E)=\dim_{\mathbb Q}E.(12)✦ For each p∈ℛn(E)p\in\mathcal R_n(E) with n>2n>2, there is a Lie polynomial qq satisfying (13): q∈Hn(ℤ),βℚ(q)∈E,β𝔽2(q)=p.q\in H_n(\mathbb Z),\qquad \beta_{\mathbb Q}(q)\in E,\qquad \beta_{\mathbb F_2}(q)=p.(13)✦

Let SS be a commutative ring and q∈Lℤq\in L_{\mathbb Z} have rational image βℚ(q)∈W\beta_{\mathbb Q}(q)\in W. Coefficient change gives (14): βS(q)∈𝒟S.\beta_S(q)\in\mathcal D_S.(14)✦ If this rational image is homogeneous of degree n>2n>2, its special defect is (15): sS(βS(q))=0.s_S(\beta_S(q))=0.(15)✦ Let q∈L𝔽2q\in L_{\mathbb F_2} have vanishing antisymmetry and special defects, and put z=x+yz=x+y. The deletion identity is (16): [x,q(x,y)+q(z,y)]+[y,q(z,y)]=0.[x,q(x,y)+q(z,y)]+[y,q(z,y)]=0.(16)✦

In the paper, the result reads: for R=ℤ(2)R=\mathbb Z_{(2)} and Kn=Lie⁡R⟨x,y⟩nK_n=\operatorname{Lie}_R\langle x,y\rangle_n, the intersection Λn=Wn∩Kn\Lambda_n=W_n\cap K_n is a saturated finite free RR-submodule whose reduction embeds in Kn/2KnK_n/2K_n with image Λ¯n\overline\Lambda_n of 𝔽2\mathbb F_2-dimension dim⁡ℚWn\dim_{\mathbb Q}W_n, every rational Vn⊂WnV_n\subset W_n gives Mn=Vn∩KnM_n=V_n\cap K_n with embedded reduction M¯n⊂Λ¯n\overline M_n\subset\overline\Lambda_n of 𝔽2\mathbb F_2-dimension dim⁡ℚVn\dim_{\mathbb Q}V_n, and the reduced elements satisfy all three Drinfeld equations and, for n>2n>2, the special identity, its linearization and all formal coefficient consequences, including, for each reduced element ψ¯\overline\psi and z=x+yz=x+y, the identity [x,ψ¯(x,y)+ψ¯(z,y)]+[y,ψ¯(z,y)]=0[x,\overline\psi(x,y)+\overline\psi(z,y)]+[y,\overline\psi(z,y)]=0.

What Is Checked

The statements concern lattices in ℤι\mathbb Z^\iota for finite ι\iota and their field scalar extensions, rational homogeneous subspaces E⊂WnE\subset W_n, and coefficient change from ℤ\mathbb Z to commutative rings. The natural inclusion of WnW_n takes values in Hn(ℚ)H_n(\mathbb Q), and the homogeneous inclusion of the reduced solution space 𝒮n\mathcal S_n takes values in Hn(𝔽2)H_n(\mathbb F_2)✦; these inclusions supply the homogeneous solution coordinate maps (4).