Definition Homogeneous maps and grading shifts
✓ definition formalized

Homogeneous maps and grading shifts

For graded \(A\)-modules \(M=\bigoplus_n M_n\) and \(N=\bigoplus_n N_n\), a homogeneous map of degree zero is an \(A\)-linear map \(f:M\to N\) satisfying \(f(M_n)\subseteq N_n\) for every integer \(n\). These are two pieces of data: the linear map and its grading-preservation condition.

Maps and functionals of an integer degree

The dual module \(M^*=\operatorname{Hom}_A(M,A)\) consists of the \(A\)-linear functionals on \(M\), with addition and scalar multiplication defined pointwise.

More generally, an \(A\)-linear map has degree \(r\in\mathbb Z\) when \[ f(M_n)\subseteq N_{n+r}\qquad(n\in\mathbb Z). \] A homogeneous functional of degree \(r\) is an \(A\)-linear map \(\lambda:M\to A\) satisfying \[ \lambda(M_n)\subseteq A_{n+r}. \] Composition adds degrees: if \(f\) has degree \(r\) and \(g\) degree \(s\), then \(g(f(M_n))\subseteq P_{n+r+s}\).

For a basis vector of degree \(d\), its coordinate functional has degree \(-d\): on a degree-\(n\) vector it reads a coefficient of degree \(n-d\). The detailed coefficient argument is in Degrees of coefficients and dual basis functionals.

Positive shifts of nonnegative module data

For \(d\in\mathbb N\), the positive shift \(M(d)\) has the same underlying module and \[ M(d)_n=M_{n-d}. \] An element of original degree \(r\) has degree \(r+d\) after this shift. For example, the element \(1\) in \(A(2)\) has degree two. This sign convention gives \(\operatorname{ch}(M(d))=q^d\operatorname{ch}(M)\) when \(M\) is graded free.

The shifted map \(f(d):M(d)\to N(d)\) is the same linear map: a vector in degree \(n\) lies in \(M_{n-d}\), whose image lies in \(N_{n-d}\). Restricting to \(d\ge0\) preserves the nonnegative grading required by the module-data object. Degree-shifting linear maps appearing in pairing arguments are stated separately with their explicit integer degree.

Identity, composition and shifted components

The identity map preserves every piece. If \(f:M\to N\) and \(g:N\to P\) have degree zero, their composite does too: \(g(f(M_n))\subseteq g(N_n)\subseteq P_n\). These are the source's Hom.id and Hom.comp operations.

The shifted component maps are \[ \pi_n^{M(d)}(m)=\pi_{n-d}^M(m). \] The finite set of original component degrees is translated by \(d\), giving a finite shifted decomposition. Projection identities and uniqueness are inherited from \(M\).

For \(a\in A_i\) and \(m\in M(d)_j=M_{j-d}\), the product lies in \(M_{i+j-d}=M(d)_{i+j}\). If \(n<0\) and \(d\ge0\), then \(n-d<0\), proving nonnegativity of the shift. Its underlying module and finite generating family stay the same.

The component-transport theorem explains homogeneous preimages, degree preservation by inverse bijections, and the induced grading on kernels and images.

References

The precise Lean sources are listed below.

Lean formalization

The fields map and graded specify the A-linear map and its degree-zero preservation.

OAI.KLInvariance.Graded.ModuleData.Hom: GradedModules.lean:1073.

OAI.KLInvariance.Graded.ModuleData.shift: GradedModules.lean:1244.

OAI.KLInvariance.Graded.ModuleData.Hom.shift: GradedModules.lean:1260.

OAI.KLInvariance.Graded.ModuleData.Hom.id: GradedModules.lean:1153.

OAI.KLInvariance.Graded.ModuleData.Hom.comp: GradedModules.lean:1148.

OAI.KLInvariance.Graded.shift_isInternal: GradedModules.lean:1226.

Used by