Theorem Rigidity of a reciprocal scaled difference
✓ theorem proved

Rigidity of a reciprocal scaled difference

Let \(P,P'\in\mathbb Z[q]\) satisfy \[ 2\,\operatorname{natDegree}(P)<d,\qquad 2\,\operatorname{natDegree}(P')<d. \] Assume that, for every real \(0<t<1\), \[ S_d(P;t)-S_d(P';t) =S_d(P;t^{-1})-S_d(P';t^{-1}). \] Then \(P=P'\).

Proof

Step 1. Choose a square root for each polynomial parameter

Fix any real \(q\) with \(0<q<1\). Put \(t=\sqrt q\). Then \(t>0\), \(t^2=q\), and \(t<1\): otherwise \(t^2\ge1\), contradicting \(q<1\).

Also \((t^{-1})^2=(t^2)^{-1}=q^{-1}\). The assumed scaled identity therefore applies to this \(t\).

Step 2. Expand the two scaled differences

Write \(\Delta(q)=P(q)-P'(q)\). Linearity of the scaling gives \[ S_d(P;t)-S_d(P';t)=t^{-d}\Delta(q), \] \[ S_d(P;t^{-1})-S_d(P';t^{-1}) =(t^{-1})^{-d}\Delta(q^{-1}) =t^d\Delta(q^{-1}). \] Thus the hypothesis becomes \[ t^{-d}\Delta(q)=t^d\Delta(q^{-1}). \]

Step 3. Cancel the scale

Since \(t>0\), the number \(t^d\) is nonzero. Multiply by it: \[ \Delta(q)=t^{2d}\Delta(q^{-1}). \] Now \(t^{2d}=(t^2)^d=q^d\). Therefore \[ \Delta(q)=q^d\Delta(q^{-1}) \] for every \(0<q<1\).

Step 4. Apply reciprocal-difference rigidity

The two original strict midpoint bounds are precisely the hypotheses of the integral reciprocal-difference theorem. The identity from Step 3 supplies its evaluation condition, so it gives \(P=P'\).

For constant polynomials and \(d=1\), the scaled difference is \((P-P')t^{-1}\), and its inverted value is \((P-P')t\). Their equality at \(t=1/2\) forces the constant difference to vanish.

References

The local source declarations below give the construction and proof.

Lean formalization

OAI.KLInvariance.Comparison.eq_of_scaled_difference_reciprocal: SimultaneousInvariance.lean:328.