Definition The ordinary and reflected tail maps of a tensor word
✓ definition formalized

The ordinary and reflected tail maps of a tensor word

Let \(C\to A\) be an algebra map between commutative rings, let \(N\) be an \(A\)-module with its restricted \(C\)-action, and let \(\tau:A\to A\) be a ring automorphism fixing the image of \(C\).

There are two \(A\)-linear maps \[ \mu:A\otimes_C N\to N,\qquad \mu(a\otimes n)=an, \] and \[ \nu:A\otimes_C N\to{}^\tau N,\qquad \nu(a\otimes n)=\tau(a)n. \] The second formula denotes the underlying vector in the twisted module, whose action is \(r\star n=\tau(r)n\). These are the source's tensorFirst and tensorSecond.

Tensor relations and the scalar twist

For \(c\in C\), the ordinary formula sends both \(ac\otimes n\) and \(a\otimes cn\) to \(acn\). The reflected formula sends them to \[ \tau(ac)n=\tau(a)c n, \] because \(\tau(c)=c\). Both maps are therefore well defined on the tensor product.

The first map is \(A\)-linear by ordinary multiplication. For the second, \[ \nu(r(a\otimes n))=\tau(ra)n =\tau(r)\tau(a)n=r\star\nu(a\otimes n), \] which proves linearity into its twisted target.

If \(h:N\to P\) is \(A\)-linear, the same underlying formula \(n\mapsto h(n)\) defines \({}^\tau h:{}^\tau N\to{}^\tau P\): the equality \(h(\tau(r)n)=\tau(r)h(n)\) gives the required scalar law.

Applying the maps to a word

For a finite-generator Coxeter system, let \(A\) be its symmetric coefficient algebra, fix a simple reflection \(s\), and put \(C=A^s\), \(\tau=\chi_s\), \(N=M_{\mathbf t}\). Then \[ \mu:M_{s\mathbf t}\to M_{\mathbf t},\qquad \nu:M_{s\mathbf t}\to{}^\tau M_{\mathbf t} \] are bottHeadTail and bottHeadReflectedTail.

For a finite tensor sum \(m=\sum_j a_j\otimes n_j\), their underlying vectors are \[ \mu(m)=\sum_j a_jn_j,\qquad \nu(m)=\sum_j\tau(a_j)n_j. \] In particular, both send \(1\otimes n\) to the underlying vector \(n\).

The local pair of tail projections

For each group weight \(x\), define \[ P_x(m)= \bigl(\pi_x(\mu(m)),\ \pi_{sx}(\nu(m))\bigr) \in M_{\mathbf t}^x\oplus{}^\tau M_{\mathbf t}^{sx}. \] The second projection here is the twisted map on the underlying vector of \(\nu(m)\). Hence \(P_x\) is \(A\)-linear. It is bottHeadLocalPair.

For \(1\otimes n\), it is the common-preimage pair \[ P_x(1\otimes n)=(\pi_x(n),\pi_{sx}(n)). \] The local-pair theorem shows that \(P_x(m)\) always lies in the head pair lattice and represents the new weight-stalk projection.

References

The local source declarations below give the construction and proof.

Lean formalization

OAI.KLInvariance.ReflectionPairTranslation.tensorFirst: BottSamelson.lean:706.

OAI.KLInvariance.ReflectionPairTranslation.tensorFirst_tmul: BottSamelson.lean:710.

OAI.KLInvariance.ReflectionPairTranslation.tensorSecond: BottSamelson.lean:713.

OAI.KLInvariance.ReflectionPairTranslation.tensorSecond_tmul: BottSamelson.lean:726.

OAI.KLInvariance.ReflectionPairTranslation.twistMap: BottSamelson.lean:729.

OAI.KLInvariance.TitsSpace.bottHeadTail: BottSamelson.lean:750.

OAI.KLInvariance.TitsSpace.bottHeadReflectedTail: BottSamelson.lean:754.

OAI.KLInvariance.TitsSpace.bottHeadTail_seed: BottSamelson.lean:760.

OAI.KLInvariance.TitsSpace.bottHeadReflectedTail_seed: BottSamelson.lean:764.

OAI.KLInvariance.TitsSpace.bottHeadLocalPair: BottSamelson.lean:769.

OAI.KLInvariance.TitsSpace.bottHeadLocalPair_seed: BottSamelson.lean:777.

Used by