Theorem Invariant and anti-invariant reflection coordinates
✓ theorem proved

Invariant and anti-invariant reflection coordinates

Let \(k\) be a characteristic-zero field, \(A\) a commutative domain over \(k\), and \(\tau\) a \(k\)-algebra involution. Suppose \[ \alpha\ne0,\qquad \tau(\alpha)=-\alpha,\qquad \alpha\mid a-\tau(a)\quad(a\in A). \] Put \(R=A^\tau\). Every \(a\in A\) has a unique expression \[ a=a_0+\alpha a_1,\qquad a_0,a_1\in R. \] Consequently the expansion \(E:R^2\to A\) is an \(R\)-linear isomorphism; its inverse is the reflection coordinate map \(C\). The elements \(1,\alpha\) form an \(R\)-basis.

Proof

Step 1. Show that the divided difference is invariant

Choose \(d\in A\) with \[ a-\tau(a)=\alpha d. \] Applying \(\tau\) and using \(\tau^2=1\) gives \[ \tau(a)-a=-\alpha\,\tau(d). \] Negating the original equality gives the same left side as \(-\alpha d\). Thus \(\alpha\tau(d)=\alpha d\). Since \(A\) is a domain and \(\alpha\ne0\), cancellation gives \(\tau(d)=d\).

Step 2. Construct both coordinates

Let \(h\) be the image of \(1/2\in k\) in \(A\). Characteristic zero gives \(2h=1\). Set \[ a_0=h(a+\tau(a)),\qquad a_1=hd. \] The automorphism fixes \(h\), and it exchanges \(a\) with \(\tau(a)\). Hence \(\tau(a_0)=a_0\). Step 1 gives \(\tau(a_1)=a_1\). Both coordinates therefore lie in \(R\).

Their expansion is \[ a_0+\alpha a_1 =h(a+\tau(a))+h(a-\tau(a))=2ha=a. \] This proves surjectivity of \(E\).

Step 3. Prove uniqueness

Suppose \(a_0+\alpha a_1=b_0+\alpha b_1\), with all four coefficients invariant. Applying \(\tau\) yields \[ a_0-\alpha a_1=b_0-\alpha b_1. \] Subtracting gives \(2\alpha(a_1-b_1)=0\). The scalar map \(k\to A\) is injective because \(k\) is a field and \(A\) is a nonzero algebra. Thus \(2\ne0\) in \(A\), and the domain property gives \(2\alpha\ne0\). It follows that \(a_1=b_1\), and the original equality then gives \(a_0=b_0\).

Step 4. Obtain the linear coordinates and the basis

The expansion map is \(R\)-linear by its definition. Its bijectivity gives the inverse \(R\)-linear map \[ C:A\longrightarrow R^2,\qquad C(a)=(a_0,a_1). \] The source's two inverse identities are \[ E(C(a))=a,\qquad C(E(c_0,c_1))=(c_0,c_1). \] Transporting the standard basis of \(R^2\) through \(E\) gives \(E(1,0)=1\) and \(E(0,1)=\alpha\). These are the source's reflectionBasis vectors.

Step 5. Identify the two eigenspaces

For \(a=a_0+\alpha a_1\), one has \(\tau(a)=a_0-\alpha a_1\). Thus \(\tau(a)=a\) forces \(2\alpha a_1=0\), giving \(a_1=0\). Likewise \(\tau(a)=-a\) forces \(2a_0=0\), giving \(a_0=0\).

The invariant summand is therefore \(R\), and the anti-invariant summand is \(\alpha R\). This explains the direct-sum expression \(A=R\oplus\alpha R\).

References

The local source declarations below give the construction and proof.

Lean formalization

OAI.KLInvariance.ReflectionFrobenius.quotient_fixed: Decomposition.lean:1462.

OAI.KLInvariance.ReflectionFrobenius.expand_injective: Decomposition.lean:1469.

OAI.KLInvariance.ReflectionFrobenius.expand_surjective: Decomposition.lean:1486.

OAI.KLInvariance.ReflectionFrobenius.coordinates: Decomposition.lean:1504.

OAI.KLInvariance.ReflectionFrobenius.expand_coordinates: Decomposition.lean:1508.

OAI.KLInvariance.ReflectionFrobenius.coordinates_expand: Decomposition.lean:1512.

OAI.KLInvariance.ReflectionFrobenius.reflectionBasis: Frobenius.lean:522.

OAI.KLInvariance.ReflectionFrobenius.reflectionBasis_zero: Frobenius.lean:525.

OAI.KLInvariance.ReflectionFrobenius.reflectionBasis_one: Frobenius.lean:529.

Used by