Definition The reflected two-stalk coordinates of a longer word
✓ definition formalized

The reflected two-stalk coordinates of a longer word

Fix a finite-generator Coxeter system, a word \(\mathbf t\), a simple reflection \(s\), and a group element \(x\). Let \(\tau=\chi_s\) on the symmetric coefficient algebra \(A\), and write \(\alpha=\alpha_s\).

Use the two tail stalks \[ X=M_{\mathbf t}^x,\qquad Y=M_{\mathbf t}^{sx}, \qquad f=\pi_x,\quad g=\pi_{sx}. \] The reflected pair lattice is \[ L_x=\{(f(n),g(n)):n\in M_{\mathbf t}\} \subseteq X\oplus{}^\tau Y, \] with action \(a\cdot(p,q)=(ap,\tau(a)q)\).

The source's bottHeadPairLattice instantiates the general twisted pair lattice. Its two hypotheses come from \[ \alpha\mid\tau(a)-a \] and the reflection-label annihilation theorem, which supplies a common preimage for every pair \((0,\alpha q)\).

Splitting the new weight block by its first bit

A subexpression \((0,\eta)\) of \(s\mathbf t\) has weight \(x\) exactly when the tail weight is \(x\). A subexpression \((1,\eta)\) has weight \(x\) exactly when the tail weight is \(sx\), because \[ s\,w_{\mathbf t}(\eta)=x \quad\Longleftrightarrow\quad w_{\mathbf t}(\eta)=sx. \] Thus the coordinate set for the new \(x\)-stalk splits into a zero-bit copy of \(J_x\) and a one-bit copy of \(J_{sx}\) for the tail.

The coordinate map

For \((p,q)\in X\oplus{}^\tau Y\), define \[ C_x(p,q)_{(0,\eta)}=p_\eta,\qquad C_x(p,q)_{(1,\eta)}=\tau(q_\eta). \] Its target is the full coordinate module for the new \(x\)-weight block. This is bottHeadPairCoordinates.

The scalar twist makes \(C_x\) \(A\)-linear. On the second block, \[ \tau((a\cdot q)_\eta) =\tau(\tau(a)q_\eta) =a\,\tau(q_\eta), \] while the first block has ordinary scalar linearity.

For a pair with common preimage \(n\), \[ C_x(f(n),g(n)) =E_{s\mathbf t,x}(1\otimes n). \] This equality follows directly from the two evaluation formulas. The head-pair equivalence theorem proves that restricting \(C_x\) to \(L_x\) gives exactly the new image stalk.

References

The local source declarations below give the construction and proof.

Lean formalization

OAI.KLInvariance.TitsSpace.bottWeight_tail_zero: BottSamelson.lean:99.

OAI.KLInvariance.TitsSpace.bottWeight_tail_one: BottSamelson.lean:105.

OAI.KLInvariance.TitsSpace.bottHeadPairCoordinates: BottSamelson.lean:112.

OAI.KLInvariance.TitsSpace.simpleReverseDivisibility: BottSamelson.lean:156.

OAI.KLInvariance.TitsSpace.bottHeadPairLattice: BottSamelson.lean:162.

Used by