Quiver Hecke superalgebras extend Khovanov–Lauda–Rouquier algebra by adding parity to the vertex data and to the generators. Their diagrams still contain strands, dots and crossings, but exchanging odd operations can introduce signs. Graded supermodule categories then retain both an integer grading and an even–odd grading when they are used in categorification.
The useful starting point is not “put a minus sign everywhere.” One must distinguish an associative superalgebra from a supercommutative algebra, distinguish parity from homological degree, and calculate how operators act on tensor products. In particular, an odd generator need not have square zero. This guide proves that point with explicit examples and develops a two-variable odd nilHecke representation in which the defining signs can be checked directly.
The central proposition is that parity changes the rules for exchanging operations, not the requirement to verify those rules. We construct a compatible mixed Cartan example, calculate an odd divided-difference operator, prove its square is zero, and show how parity shifts become a formal parameter in a Grothendieck group. The final sections explain which quantum-group and quantum-supergroup statements are justified by the cited constructions.
Scope. We work over C, so two is invertible. The principal Cartan-data examples use the anisotropic framework with diagonal Cartan entries two and the parity compatibility described below. This is not an assertion that the same presentation covers every Lie superalgebra with isotropic odd simple roots. Kang–Kashiwara–Tsuchioka, Hill–Wang and Kang–Kashiwara–Oh supply the precise algebraic and categorical frameworks. [1] [3] [4]
Your 50-second route
New to superalgebra? Start with parity and tensor signs. Coming from ordinary KLR algebras? Go to the Cartan and generator data. Ready for a calculation? Use the odd nilHecke laboratory. Studying categorification? Read grading, parity shifts and cyclotomic quotients. Test the distinctions in the worked practice.
Open the complete learning map
1–7: even–odd spaces, superalgebras and tensor signs
8–13: Cartan data, generators and a mixed example
14–20: skew polynomials and odd divided differences
21–28: Grothendieck groups, induction, cyclotomic quotients and scope
Practice and worked answers
Primary sources and companion pages
Teaching guide
The ordinary prerequisite is KLR and Quiver Hecke Algebras. Further background is available in Kac–Moody Algebras, Crystal Bases and Categorification.
1. A super vector space has two parts
A super vector space is a direct sum V=V₀⊕V₁. V₀ is the even part and V₁ is the odd part. A homogeneous vector lies in one part and has parity |v| equal to zero or one. The sum of an even and an odd vector is a perfectly valid vector, but it does not have one homogeneous parity.
If dim V₀=a and dim V₁=b, write the superdimension pair as a|b. The ordinary total dimension is a+b. The numerical superdimension is a−b. These are different measurements of the same space. A space of dimension 1|1 has total dimension two and superdimension zero; it is not the zero space.
The terminology is mathematical. Calling a vector odd does not make it a physical fermion, and a super vector space alone does not specify a physical system. The present guide uses parity to organize algebraic signs and representation categories.
2. Associative superalgebra is not the same as supercommutative algebra
An associative superalgebra A=A₀⊕A₁ satisfies A_iA_j⊂A_(i+j mod 2). This only says that multiplication respects parity. Supercommutativity is the additional rule ab=(−1)^(|a||b|)ba for homogeneous elements. The extra rule is not part of the definition of every superalgebra.
In a supercommutative algebra over C, an odd element x satisfies x²=−x² and hence x²=0. In a general associative superalgebra, there is no such deduction. For example, the ordinary polynomial algebra C[x] can be graded by declaring x odd and parity equal to exponent modulo two. It is an associative superalgebra with x² nonzero, but it is not supercommutative.
This distinction is crucial in odd nilHecke theory. Distinct odd polynomial generators anticommute, but their squares are not set to zero. Replacing that algebra by an exterior algebra would remove genuine basis elements and change the entire representation problem.
3. An odd invertible operator
Take V with even basis vector e and odd basis vector f. Define C(e)=f and C(f)=e. Then C is odd because it exchanges the parity spaces. Its matrix in the basis (e,f) is [[0,1],[1,0]], and C²=I.
This gives a module for the associative superalgebra generated by one odd element c satisfying c²=1. The example is a small Clifford-type model. Its existence directly disproves the statement that every odd operator is nilpotent.
The ordinary commutator [C,C] is zero, while the supercommutator [C,C]_s=CC−(−1)CC=2I. Thus the choice of commutator convention matters. A formula written with a bracket in a superalgebra must specify whether it is an ordinary commutator or a parity-adjusted one.
4. The super flip is a signed interchange
For homogeneous vectors, the super flip sends v⊗w to (−1)^(|v||w|)w⊗v. Interchanging two odd vectors contributes a minus sign; interchanging an even vector with anything contributes no sign. Applying the flip twice gives the identity because the two signs multiply to one.
For the 1|1 space above, e⊗e is fixed, e⊗f and f⊗e are exchanged, and f⊗f is sent to its negative. The tensor product has even part spanned by e⊗e and f⊗f, and odd part spanned by e⊗f and f⊗e. Its dimension is 2|2 and its total dimension is four.
The vector f⊗f is nonzero in the tensor product. The minus sign from the flip does not force it to vanish. It would vanish only after imposing an additional quotient relation that identifies the vector with its signed flip. Tensor products and symmetric or exterior quotients must not be conflated.
5. Tensoring operators introduces a different sign
For homogeneous linear maps A and B, the super tensor action is (A⊗B)(v⊗w)=(−1)^(|B||v|)Av⊗Bw. The sign occurs because B passes the first input v. Consequently, composing two tensor operators gives (A⊗B)(C⊗D)=(−1)^(|B||C|)AC⊗BD.
Let C be the odd involution from the previous example. On V⊗V, the operators C⊗I and I⊗C anticommute. Applied to e⊗e, the first order gives f⊗f and the reverse order gives −f⊗f. Therefore their ordinary commutator is twice one composition, while their supercommutator is zero.
This calculation explains why distant odd operations in diagrammatics can acquire signs even when they act in separate tensor positions. The sign is not an arbitrary decoration attached to a crossing. It follows from the tensor convention and the parities of the operations that pass one another.
6. Integer grading and parity are independent
A graded super vector space has pieces V_(d,ε), where d∈Z is an integer degree and ε∈{0,1} is parity. An even vector may have odd integer degree, and an odd vector may have even integer degree. One does not determine the other unless a separate convention imposes a relationship.
For example, take an even vector of degree zero, an odd vector of degree one, and one even and one odd vector of degree two. Record this by D(q,π)=1+πq+(1+π)q², with π²=1. Setting π=1 gives the ordinary graded dimension 1+q+2q². Setting π=−1 gives the signed graded dimension 1−q.
At q=1, the total dimension is four while the superdimension is zero. Neither computation identifies the space with zero. This example also warns against confusing parity signs with the alternating signs of an Euler characteristic, which come from homological degree unless an explicit identification is made.
7. Characteristic two changes the discussion
The deduction x²=0 from supercommutativity used 2x²=0 and division by two. The difference between commuting and anticommuting also depends on −1 differing from 1. Over characteristic two, those signs collapse and the usual formulas require a different interpretation.
Our calculations therefore stay over C. This is not a claim that no useful super representation theory exists in characteristic two; it is a boundary on the formulas being used here. A theorem stated over a coefficient ring with two invertible should not be applied after silently changing that ring.
Field assumptions belong beside parity assumptions in a working notebook. Before comparing two presentations, check both. A sign discrepancy may arise from a convention change, but in characteristic two it may reflect a genuinely different algebraic setting.
8. Parity must be compatible with the Cartan datum
In the anisotropic Cartan framework used here, each vertex i has a parity p(i). The generalized Cartan matrix has diagonal entries two, and the required compatibility includes even entries in a row corresponding to an odd vertex. One cannot arbitrarily mark a vertex odd in every ordinary Dynkin diagram and assume the same superalgebra construction applies. [1] [3]
For instance, the ordinary A₂ matrix [[2,−1],[−1,2]] does not satisfy that even-row condition if either vertex is declared odd. The failure is visible before writing a single dot or crossing relation. A diagram with parity is additional structured data, not merely a diagram with a changed color.
This boundary is especially important when comparing with Lie superalgebras having isotropic odd simple roots. Such roots have a different Cartan behavior. The references cited for the present categorification have specific scopes; they do not justify replacing every super root system by this same matrix grammar.
9. A compatible two-vertex example
Take the matrix C=[[2,−2],[−1,2]]. Declare vertex 1 odd and vertex 2 even. The odd row has entries two and minus two, both even. Multiplying on the left by D=diag(1,2) gives DC=[[2,−2],[−2,4]], which is symmetric.
The symmetrizing numbers d₁=1 and d₂=2 assign integer dot degrees 2d₁=2 and 2d₂=4. The first dot is odd despite having even integer degree two. The second is even with integer degree four. This is a concrete reminder that the two gradings must be recorded separately.
The mixed crossing has parity p(1)p(2)=0 and degree −d₁c₁₂=2. A crossing between two odd copies of vertex 1 instead has odd parity and negative integer degree −2. Negative degree of an operator is allowed; it records how the operator changes the module grading, not a negative number of vectors.
10. Sequence idempotents organize the algebra
Fix a positive root combination β and consider all sequences ν=(ν₁,…,ν_n) of vertices whose simple-root sum is β. The idempotent e(ν) records one such sequence. Different sequence idempotents are orthogonal, and their sum is the identity of the corresponding algebra.
A dot x_r acts at position r and has parity p(ν_r) on e(ν). A crossing τ_r exchanges adjacent positions r and r+1 and has parity p(ν_r)p(ν_(r+1)). The idempotent bookkeeping identifies the source and target color sequences of a diagram.
For the sequence (1,2), the first dot is odd, the second is even, and the crossing is even. For (1,1), both dots and the crossing are odd. The parity of a named generator can therefore depend on the sequence component on which it acts. Suppressing the idempotent can hide that information.
11. Polynomial relations must respect parity and degree
The defining data include polynomials Q_ij(u,v) controlling crossing squares. They satisfy Q_ii=0 and compatibility under exchanging i,j and u,v. Variables associated with odd vertices must enter with the specified evenness conditions. These restrictions are part of the definition in Kang–Kashiwara–Tsuchioka. [1]
For our mixed example, choose Q₁₂(u,v)=u²−v and Q₂₁(u,v)=v²−u. Then Q₂₁(v,u)=Q₁₂(u,v). In Q₁₂, the first variable belongs to the odd vertex and appears only through u². In Q₂₁, the second variable belongs to the odd vertex and appears only through v².
The weighted degrees also agree. In Q₁₂, u has degree two and v has degree four, so both terms have degree four. That equals twice the mixed-crossing degree. This is an explicit local compatibility check; it does not replace the remaining braid and dot-crossing relations of the full algebra.
12. Distinct odd dots anticommute, but a dot may square nontrivially
For two different positions r and s, the dot relation includes x_rx_s e(ν)=(−1)^(p(ν_r)p(ν_s))x_sx_r e(ν). On an all-odd two-strand sequence this says x₁x₂=−x₂x₁. It does not say x₁²=0 or x₂²=0 because the relation is for different positions.
Indeed, x₁² commutes with x₂: moving x₂ past two copies of x₁ contributes two minus signs. Explicitly, x₁²x₂=x₁(−x₂x₁)=−(−x₂x₁)x₁=x₂x₁². The square is even and can be a nonzero central polynomial in this small skew-polynomial sector.
This local calculation explains why the odd polynomial algebra has infinitely many powers of each variable. Exterior-algebra intuition would incorrectly discard them. The algebra is signed, but not globally supercommutative.
13. The equal-odd-color crossing leads to odd nilHecke relations
On two strands of the same odd color, the crossing τ satisfies τ²=0 and mixed relations τx₂+x₁τ=1 and x₂τ+τx₁=1, in the convention used here. The dots satisfy x₁x₂=−x₂x₁. These are the local odd nilHecke relations that will be realized explicitly in the next section. [1] [2]
The plus signs in the mixed relations are not obtained by making every ordinary commutator negative. They arise from the precise parities and crossing conventions. A reliable method is to construct an operator satisfying the relations and test it on a basis.
For more strands, one also needs distant-crossing signs, braid relations and their color-dependent corrections. The two-strand calculation is a local laboratory, not a claim that four displayed equations present the entire quiver Hecke superalgebra for all color sequences.
14. Build the skew-polynomial space
Let P be the associative algebra generated by odd x₁,x₂ with x₁x₂=−x₂x₁. Put monomials into the order x₁ᵃx₂ᵇ, where a,b are nonnegative integers. Multiplying two ordered monomials gives (x₁ᵃx₂ᵇ)(x₁ᶜx₂ᵈ)=(−1)^(bc)x₁^(a+c)x₂^(b+d).
The sign counts how many times a copy of x₂ from the first factor passes a copy of x₁ from the second. For example, (x₁x₂)(x₁²x₂³)=x₁³x₂⁴ because two crossings contribute a positive sign. But (x₁x₂)(x₁x₂³)=−x₁²x₂⁴ because only one crossing is needed.
For computational bookkeeping in this section, total polynomial degree means a+b. The KLR integer grading can instead assign each same-color dot degree two; then all polynomial degrees are doubled and the divided-difference operator below has degree minus two. The parity remains a+b modulo two in either convention.
15. A signed swap automorphism
Define s(x₁)=−x₂ and s(x₂)=−x₁, extending s multiplicatively and linearly. Applying s twice fixes both generators, so s²=1. It respects the skew relation because s(x₁x₂+x₂x₁)=x₂x₁+x₁x₂=0.
This s is an algebra automorphism. It is not the same notation as an ordinary unsigned permutation of commuting variables. For example, s(x₁x₂)=x₂x₁=−x₁x₂, whereas s(x₁²)=x₂².
The distinction matters in the Leibniz rule. The operator we are about to construct uses s(f), not merely a sign times f. It is a twisted divided difference, not an ordinary derivative with a parity sign added by hand.
16. Define the odd divided difference
Define ∂(1)=0, ∂(x₁)=∂(x₂)=1, and extend by the twisted rule ∂(fg)=∂(f)g+s(f)∂(g). To check that the rule descends to P, apply it to the relation: ∂(x₁x₂)=x₂−x₂=0 and ∂(x₂x₁)=x₁−x₁=0. Thus ∂ annihilates x₁x₂+x₂x₁, and s preserves that relation.
The operator reverses parity and lowers total polynomial degree by one. Its definition is a concrete version of the odd divided-difference construction developed by Ellis, Khovanov and Lauda. [2]
There is no division by x₁−x₂ in this definition. The recursive twisted Leibniz rule specifies its action directly. This is helpful for checking signs because every step uses either a known generator value, a multiplication in P or the signed swap s.
17. Calculate low-degree monomials
For x₁², the rule gives ∂(x₁²)=x₁−x₂. Likewise ∂(x₂²)=x₂−x₁. We already found ∂(x₁x₂)=0. These three formulas completely determine the operator on the degree-two ordered monomial basis.
For x₁³, split the product as x₁²·x₁. Then ∂(x₁³)=(x₁−x₂)x₁+x₂². Reordering x₂x₁=−x₁x₂ gives ∂(x₁³)=x₁²+x₁x₂+x₂². Applying ∂ once more gives (x₁−x₂)+0+(x₂−x₁)=0.
For x₁²x₂, the result is (x₁−x₂)x₂+x₂²=x₁x₂. For x₁x₂², it is x₂²−x₂(x₂−x₁)=x₂x₁=−x₁x₂. The opposite signs in these two examples expose the noncommutativity of the polynomial variables.
18. Prove the mixed nilHecke relations as operator identities
Write X₁ and X₂ for left multiplication by x₁ and x₂. For every polynomial f, the twisted rule gives ∂(x₁f)=f−x₂∂f. Therefore ∂X₁+X₂∂=I. Similarly, ∂(x₂f)=f−x₁∂f, so ∂X₂+X₁∂=I.
These are identities on every element of P, not just on the first few monomials. They realize the mixed relations for τ=∂. The dot relation X₁X₂=−X₂X₁ follows directly from multiplication in P.
For a numerical-style check, apply ∂X₂+X₁∂ to x₁. The first term is ∂(x₂x₁)=0, while the second is x₁∂(x₁)=x₁. The result is x₁, as required by the identity. This bounded check illustrates the general proof rather than replacing it.
19. Prove that ∂²=0
First observe that ∂s=−s∂ on the generators. For example, ∂s(x₁)=∂(−x₂)=−1, while −s∂(x₁)=−1. The same holds for x₂ and for the unit. The twisted Leibniz rule then extends the identity to products, so it holds throughout P.
Apply ∂ twice to fg. Expanding produces ∂²(f)g+s(∂f)∂g+∂s(f)∂g+s²(f)∂²(g). The middle two terms cancel because ∂s=−s∂, and s²=1. Thus ∂²(fg)=∂²(f)g+f∂²(g).
The operator ∂² is therefore an ordinary derivation that vanishes on the unit and both generators. It vanishes on every polynomial. This proves ∂²=0 and completes the two-strand odd nilHecke representation check. Notice the distinction: the odd crossing squares to zero because of its specific relation and operator construction, while the odd dots do not.
20. What the local model proves, and what it does not
We have constructed actual operators satisfying the two-strand equal-odd-color dot, crossing-square and mixed relations. The proof used a concrete skew-polynomial algebra and a twisted derivation. It explains exactly where the signs occur and why the square-zero statement belongs to ∂ rather than to every odd generator.
We have not thereby verified all color-dependent braid relations for an arbitrary quiver. Nor have we proved a categorification theorem or identified every simple module. Those are additional structural statements supplied by the primary references.
This separation is productive. A local model gives an entry point that can be calculated from first principles. The global theory explains how many such local pieces fit together and what their module categories recover after taking Grothendieck groups.
21. Grading shift and parity shift leave different records
An integer grading shift M⟨1⟩ moves graded pieces by one and is recorded by multiplication by q in a graded Grothendieck group, according to the declared shift convention. A parity shift ΠM exchanges the even and odd parts. Applying it twice returns an isomorphic supermodule, so its formal parameter π satisfies π²=1.
One therefore keeps track of classes over a coefficient ring involving q,q⁻¹ and π with π²=1. The relation [ΠM]=π[M] does not mean [ΠM]=−[M] before a specialization is chosen. Setting π=−1 produces a signed invariant; setting π=1 forgets the parity distinction.
Hill and Wang make parity shift central to the categorification of the super sign in their anisotropic quantum-superalgebra framework. [3] The sign is thus represented by an operation on objects, not introduced as an unexplained negative coefficient.
22. A one-strand cyclotomic calculation
For one odd strand, the dot algebra is generated by an odd x with no relation forcing x²=0. Impose the homogeneous relation x³=0. The quotient has basis 1,x,x². With dot degree two, its graded superdimension is 1+πq²+q⁴.
The even part is spanned by 1 and x², while the odd part is spanned by x. Thus the quotient has dimension 2|1 and total dimension three. Multiplication by x sends 1 to x, x to x² and x² to zero. It is an odd nilpotent operator of order three, not necessarily order two.
This is the simplest local model of a cyclotomic dot truncation. In a full cyclotomic quiver Hecke superalgebra, the exponent is determined by highest-weight data and the relation is imposed on the appropriate sequence components. The one-strand example illustrates the mechanism without claiming to compute every higher-strand quotient.
23. Induction concatenates color sequences
Given root combinations β and γ, juxtaposing color sequences produces a corner associated with β+γ. The corresponding algebra maps define induction functors from modules for the smaller algebras to modules for the larger one. Restriction recovers selected sequence components.
In the super setting, tensor products of algebras and modules use the signed convention already calculated. Omitting those signs can make an alleged algebra embedding fail to preserve multiplication. The local tensor identity (A⊗B)(C⊗D)=(−1)^(|B||C|)AC⊗BD is therefore part of the induction machinery, not an optional preliminary.
Passing to Grothendieck groups turns induction into an algebra product and turns grading and parity shifts into formal coefficients. The categorification theorem identifies that algebraic structure with a specified quantum or covering algebra under its stated hypotheses. It is stronger than the observation that diagrams can be concatenated.
24. Projectives, simples and basis claims
Finitely generated projective modules and finite-dimensional simple modules provide different Grothendieck-group descriptions. Projective classes naturally encode multiplication and decomposition, while simple classes can encode dual information and crystal-like operations. Their bases should not be identified without a theorem and a normalization.
In ordinary KLR theory, canonical-basis statements already depend on characteristic and the chosen realization. Adding parity introduces further distinctions between even isomorphisms, parity shifts and supermodule equivalences. A result about projectives is not automatically a classification of all simples.
The safe question is specific: which category, which Grothendieck group, which coefficient ring, and which map to the target algebra? Once those are written, a basis claim becomes a mathematical statement that can be checked rather than a broad assertion that “the category is the quantum group.”
25. Quantum groups and quantum supergroups are not one unqualified target
Hill–Wang use spin quiver Hecke algebras to categorify a half of quantum Kac–Moody superalgebras in the anisotropic setting. Kang–Kashiwara–Oh establish supercategorification results for quantum Kac–Moody algebras and their integrable highest-weight modules using quiver Hecke superalgebras and cyclotomic quotients. The targets and treatments of parity must be read as stated in the respective papers. [3] [4]
This is not a contradiction. A category retaining parity can support different decategorified specializations or comparison constructions. Forgetting parity, identifying parity shifts, or specializing a covering parameter changes the invariant being extracted.
The lesson is the same as the example D(q,π): π=1 and π=−1 are different outputs from the same two-graded input. A guide should specify which output is intended rather than saying that every superalgebra presentation automatically categorifies every quantum supergroup.
26. Hecke–Clifford comparisons retain their qualifiers
Kang–Kashiwara–Tsuchioka introduce quiver Hecke–Clifford superalgebras and compare them with quiver Hecke superalgebras through weak Morita superequivalence. They also obtain isomorphisms with completed affine Hecke–Clifford and degenerate affine Sergeev constructions in the specified setting. [1]
The words completed and superequivalence matter. A completion allows controlled infinite expressions or isolates spectral behavior; it is not automatically the original uncompleted algebra. A supercategorical equivalence records parity information that an ordinary ungraded module comparison can forget.
Our one-generator Clifford example gives a concrete reason to preserve that information. An odd invertible operator can connect parity sectors, so forgetting which maps are even or odd changes how module structure is described. The advanced equivalence theorem is not merely an isomorphism of unsigned vector spaces.
27. What a crystal-like operation must retain
Induction and restriction by one color can be used to define operators on suitable simple-module labels. In categorification settings, these operations connect with the combinatorics of highest-weight representations. However, one must decide how labels differing by parity shift are treated and which head, socle or distinguished summand is selected.
A graph drawn from legal color additions alone does not prove a crystal theorem. It lacks the module-theoretic selection rule, weight data and compatibility identities. The same warning appeared in the quantum toroidal chapter, where legal box moves did not determine their coefficients.
The ordinary Crystal Bases guide provides the starting vocabulary. A super extension must add its parity conventions rather than pretending those conventions have no effect on the categorical objects.
28. A complete local verification checklist
Begin with the coefficient field and the allowed Cartan datum. Assign parity to each vertex and verify the compatibility conditions. Record each generator’s integer degree and parity separately. Check the Q-polynomials for symmetry, odd-variable conditions and homogeneity. Then test the dot and crossing relations on a concrete small module.
For the odd nilHecke sector, calculate ∂ on generators, a mixed monomial and a cubic monomial. Prove the mixed operator identities and ∂²=0 using the twisted Leibniz rule. For a categorical statement, name the Grothendieck group and the role of q and π. For a comparison theorem, preserve completion and superequivalence qualifiers.
These checks keep the subject from collapsing into sign memorization. Every minus sign has a source: tensor interchange, skew-polynomial reordering, a signed swap, a parity shift or a defining crossing relation. Identifying that source is more reliable than trying to remember a long formula in isolation.
Practice: diagnose the parity rule before calculating
1. A super vector space has dimension 3|2. Find its total dimension and superdimension. 2. Is an odd element’s square always zero in an associative superalgebra? 3. Compute the super flip of f⊗f when f is odd. 4. In the skew-polynomial algebra, simplify x₂³x₁².
5. Compute ∂(x₁²), ∂(x₁x₂) and ∂(x₁³). 6. Prove ∂X₂+X₁∂=I from the twisted rule. 7. Why does the argument for ∂²=0 not imply x₁²=0? 8. Verify Q₁₂(u,v)=u²−v is homogeneous for the mixed Cartan example. 9. Find the graded superdimension of C[x]/(x⁴), with odd x of integer degree two. 10. Why must a categorical theorem distinguish π=1 from π=−1?
Worked answers 1–4: parity is not disappearance
1. Total dimension is 3+2=5, while superdimension is 3−2=1. The pair 3|2 contains more information than either number separately.
2. No. Square zero follows for odd elements in a supercommutative algebra over a field where two is invertible. A general associative superalgebra only requires parity-compatible multiplication. The odd involution C with C²=I is an explicit counterexample.
3. The result is −f⊗f. The tensor itself need not vanish. A signed interchange is not the same as imposing a quotient relation identifying the tensor with its image.
4. Moving two copies of x₁ past three copies of x₂ creates six sign changes. Thus x₂³x₁²=x₁²x₂³. The sign is positive because six is even.
Worked answers 5–7: use the actual operator construction
5. The answers are x₁−x₂, zero, and x₁²+x₁x₂+x₂². Each follows from ∂(fg)=∂(f)g+s(f)∂(g), together with s(x₁)=−x₂ and skew reordering. In particular, the positive mixed term in the cubic arises because −x₂x₁=+x₁x₂.
6. For every f, ∂(x₂f)=∂(x₂)f+s(x₂)∂f=f−x₁∂f. Rearranging gives (∂X₂+X₁∂)f=f. Since f was arbitrary, this is an operator identity.
7. The proof of ∂²=0 uses the special twisted Leibniz rule and the identity ∂s=−s∂. Multiplication by x₁ is a different operator with different relations. Odd parity alone did not supply the square-zero conclusion.
Worked answers 8–10: grading and decategorification
8. The first dot variable has integer degree two, so u² has degree four. The second has degree four, so v also has degree four. The polynomial is homogeneous, and the odd variable u occurs with an even exponent. These are separate compatibility checks, both satisfied.
9. The basis is 1,x,x²,x³, with alternating parity. The graded superdimension is 1+πq²+q⁴+πq⁶. The total dimension is four and the superdimension at q=1,π=−1 is zero. The quotient is still a nonzero algebra.
10. π=1 forgets the distinction between a module and its parity shift, while π=−1 records that shift as a sign. These produce different invariants. A categorification theorem must identify the intended target and specialization rather than treating the two outputs as identical.
Primary sources and theorem scope
[1] Seok-Jin Kang, Masaki Kashiwara and Shunsuke Tsuchioka, Quiver Hecke Superalgebras. Primary source for the parity-dependent presentation, Hecke–Clifford comparison and completion statements.
[2] Alexander Ellis, Mikhail Khovanov and Aaron Lauda, The Odd nilHecke Algebra and Its Diagrammatics. The odd polynomial and divided-difference laboratory follows this framework; its displayed calculations are derived explicitly above.
[3] David Hill and Weiqiang Wang, Categorification of Quantum Kac–Moody Superalgebras, also published in Transactions of the American Mathematical Society 367 (2015). The anisotropic hypothesis and the role of parity shift are essential parts of the theorem.
[4] Seok-Jin Kang, Masaki Kashiwara and Se-jin Oh, Supercategorification of Quantum Kac–Moody Algebras and Supercategorification of Quantum Kac–Moody Algebras II. Read for the precise categories, cyclotomic quotients and target quantum algebras.
Representation Mathematics — Batch 17
Elliptic Hall Algebra: rank-degree geometry, counting and Hilbert schemes
Quantum Toroidal Algebras: currents, vector modules and Fock spaces
Affine Hecke Algebras: Iwahori vectors and explicit reducibility
Return to the BTT Mathematics Learning Hub
Teaching guide: every sign should have a reason
Begin with the odd involution on a 1|1 space. Ask whether its square must vanish, then calculate it. This immediately separates parity from nilpotence. Next calculate the super flip and two odd operators acting in different tensor positions. The learner should locate the exchanged odd objects that produce each sign.
Move to the skew-polynomial algebra and require ordered monomials before applying ∂. Compute the quadratic and cubic examples, then prove the two mixed relations and ∂²=0. Do not accept a memorized statement that “odd derivatives square to zero” as a substitute for the twisted-Leibniz proof.
Finish with the mixed Cartan example and the polynomial D(q,π). Ask which conditions come from degree, which come from parity, and which come from the chosen categorification theorem. The learner is ready to continue when those three sources of structure can be distinguished and verified independently.