Lean-certified four-point HRT results for three lattice points and one off-lattice point

Vignon Oussa · 2026

Read Lean-certified four-point HRT results for three lattice points and one off-lattice point by Vignon Oussa (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-21228.txt.

Introduction

Vignon Oussa delivers a Lean-certified theorem package for a four-point configuration: three lattice points plus one off-lattice point in the plane. The main theorem says the four linked vectors stay linearly independent for every nonzero square-integrable function when a symplectic area tops one and the coordinates are independent over the rationals. A second theorem covers the rational-coordinate case via Linnell's theorem. A full certification ledger is appended.

Catalog record

CREATORVignon Oussa
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE11459
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.21228

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Lean-certified four-point HRT results for three lattice points and one off-lattice point Vignon Oussa Abstract . We record a Lean-certified theorem package for the four-point Heil–Ramanathan–Topiwala configuration Λ = { 0 , a , b , ν } ⊂ ℝ 2 , L 0 = ℤ ​ a + ℤ ​ b , ν = r ​ a + s ​ b , \Lambda=\{0,a,b,\nu\}\subset\mathbb{R}^{2},\qquad L_{0}=\mathbb{Z}a+\mathbb{Z}b,\qquad\nu=ra+sb, with a a and b b linearly independent. The principal certified theorem states that if | σ ⁡ ( a , b ) | > 1 |\sigma(a,b)|>1 and 1 , r , s 1,r,s are linearly independent over ℚ \mathbb{Q} , then for every nonzero f ∈ L 2 ​ ( ℝ ) f\in L^{2}(\mathbb{R}) the four vectors f , π ⁡ ( a ) ​ f , π ⁡ ( b ) ​ f , π ⁡ ( ν ) ​ f f,\qquad\pi(a)f,\qquad\pi(b)f,\qquad\pi(\nu)f are linearly independent. A second certified theorem treats the rational-coordinate case r , s ∈ ℚ r,s\in\mathbb{Q} , where the configuration lies in a finer full-rank lattice and linear independence follows from Linnell’s theorem. The paper is written in standard mathematical prose. An appendix records the precise Lean certification ledger and the explicit analytic inputs used by the formal development and a download link is provided. 2020 Mathematics Subject Classification 42C15, 22E27, 03B35 1. Introduction Let T_{x}f(t)=f(t-x),\qquad M_{\omega}f(t)=e^{2\pi i\omega t}f(t),\qquad\pi(x,\omega)=M_{\omega}T_{x}, so that \pi denotes the usual projective Schrödinger representation on L^{2}(\mathbb{R}) . The Heil–Ramanathan–Topiwala (HRT) conjecture asserts that for every nonzero f\in L^{2}(\mathbb{R}) and every finite set of distinct points \Lambda\subset\mathbb{R}^{2} , the family \{\pi(\lambda)f:\lambda\in\Lambda\} is linearly independent [ 3 ] . In this paper we consider the four-point configuration \Lambda=\{0,a,b,\nu\}\subset\mathbb{R}^{2},\qquad L_{0}=\mathbb{Z}a+\mathbb{Z}b,\qquad\nu=ra+sb, where a and b are linearly independent. Our aim is to present, in conventional mathematical language, the theorem package that has been certified in Lean 4. The certified content of the paper is as follows. Status Certified statement ✓ Main theorem. If |\sigma(a,b)|>1 and 1,r,s are linearly independent over \mathbb{Q} , then for every nonzero f\in L^{2}(\mathbb{R}) the four vectors f , \pi(a)f , \pi(b)f , and \pi(\nu)f are linearly independent. ✓ Rational-coordinate theorem. If r,s\in\mathbb{Q} , then \Lambda is contained in a finer full-rank lattice; hence \{\pi(\lambda)f:\lambda\in\Lambda\} is linearly independent for every nonzero f\in L^{2}(\mathbb{R}) . ✓ Supporting infrastructure. The cyclic-subspace lemmas used in the proof of the main theorem are formalized in Lean and compiled without sorry . The proof of the principal theorem is coordinate-free. It relies on standard time-frequency input: projective commutation and strong continuity of time-frequency shifts, irreducibility of the Schrödinger representation, the three-point HRT theorem, lattice linear independence, and the standard density/incompleteness theory for Gabor systems on lattices [ 2 , 3 , 4 , 5 ] . In the formal development, these analytic ingredients are gathered into a single context object; the deduction of the certified theorems from that context is then formalized in Lean. The exact ledger is given in Appendix A . We write \sigma((x,\omega),(y,\eta))=x\eta-y\omega for the standard symplectic form on \mathbb{R}^{2} . For the lattice L_{0}=\mathbb{Z}a+\mathbb{Z}b one has \operatorname{covol}(L_{0})=|\sigma(a,b)|. This is the quantity that enters the large-covolume hypothesis in the main theorem. 2. Statement of the main theorem Theorem 2.1 (Dense large-covolume case) . Let a,b\in\mathbb{R}^{2} be linearly independent, let \nu=ra+sb , and let f\in L^{2}(\mathbb{R}) be nonzero. Assume that |\sigma(a,b)|>1\qquad\text{and}\qquad 1,r,s\ \text{are linearly independent over }\mathbb{Q}. Then the four vectors f,\qquad\pi(a)f,\qquad\pi(b)f,\qquad\pi(\nu)f are linearly independent over \mathbb{C} . This is the principal certified theorem in the Lean package. In the formal files it appears as hrt_dense_large_covolume ; the equivalent finite-family linear-independence formulation appears as hrt_dense_large_covolume_lindep . Those names are recorded in Appendix A . 3. Proof of the main theorem We begin with the lattice-cyclic subspace generated by f . Definition 3.1 . Let V=\overline{\operatorname{span}_{\mathbb{C}}\{\pi(\ell)f:\ell\in L_{0}\}}. We call V the lattice-cyclic subspace of f associated with L_{0} . Proof of Theorem 2.1 . Assume, for contradiction, that there exist coefficients c_{0},c_{1},c_{2},c_{3}\in\mathbb{C} , not all zero, such that (1) c_{0}f+c_{1}\pi(a)f+c_{2}\pi(b)f+c_{3}\pi(\nu)f=0. If c_{3}=0 , then ( 1 ) is a nontrivial linear dependence among the three vectors f , \pi(a)f , and \pi(b)f . Since 0 , a , and b are distinct and a,b are linearly independent, this contradicts the three-point HRT theorem [ 3 ] . Hence c_{3}\neq 0 . Because 0,a,b\in L_{0} , we have f\in V,\qquad\pi(a)f\in V,\qquad\pi(b)f\in V. By rearranging ( 1 ) we obtain \pi(\nu)f=-c_{3}^{-1}\bigl(c_{0}f+c_{1}\pi(a)f+c_{2}\pi(b)f\bigr)\in V. Next we claim that \pi(\nu)V\subseteq V . It is enough to verify this on the generators \pi(\ell)f with \ell\in L_{0} . By projective commutation there exists, for each \ell\in L_{0} , a nonzero scalar \kappa(\nu,\ell) such that \pi(\nu)\pi(\ell)f=\kappa(\nu,\ell)\,\pi(\ell)\pi(\nu)f. Since \pi(\nu)f\in V and V is invariant under every lattice shift \pi(\ell) , the right-hand side belongs to V . Therefore the left-hand side belongs to V as well. By linearity and closure, \pi(\nu)V\subseteq V . It follows that \pi(\ell+n\nu)V\subseteq V\qquad\text{for all }\ell\in L_{0}\text{ and }n\in\mathbb{N}. Because 1,r,s are linearly independent over \mathbb{Q} , the forward semigroup L_{0}+\mathbb{N}\nu is dense in \mathbb{R}^{2} by Kronecker’s theorem. Let z\in\mathbb{R}^{2} and h\in V . Choose a sequence z_{k}\in L_{0}+\mathbb{N}\nu with z_{k}\to z . Then \pi(z_{k})h\in V for every k , and strong continuity of z\mapsto\pi(z) implies that \pi(z_{k})h\to\pi(z)h . Since V is closed, we obtain \pi(z)h\in V . Thus \pi(z)V\subseteq V\qquad\text{for all }z\in\mathbb{R}^{2}. Now V is nonzero because f\in V and f\neq 0 . Irreducibility of the Schrödinger representation therefore forces V=L^{2}(\mathbb{R}). On the other hand, V is the closed span of the lattice Gabor orbit associated with L_{0} . Since \operatorname{covol}(L_{0})=|\sigma(a,b)|>1, standard density theory implies that this lattice Gabor orbit is incomplete in L^{2}(\mathbb{R}) [ 2 , 5 ] . Hence V\neq L^{2}(\mathbb{R}) , a contradiction. Therefore no nontrivial relation ( 1 ) can exist, and the four vectors in the statement of the theorem are linearly independent. ∎ 4. The rational-coordinate case Theorem 4.1 (Rational-coordinate case) . Let a,b\in\mathbb{R}^{2} be linearly independent, let \nu=ra+sb , and suppose that r,s\in\mathbb{Q} . Then the configuration \Lambda=\{0,a,b,\nu\} is contained in a finer full-rank lattice. Consequently, for every nonzero f\in L^{2}(\mathbb{R}) , the family \{\pi(\lambda)f:\lambda\in\Lambda\} is linearly independent. Proof. Choose N\in\mathbb{N} such that Nr,Ns\in\mathbb{Z} . Then \nu\in\frac{1}{N}\mathbb{Z}a+\frac{1}{N}\mathbb{Z}b. Since 0 , a , and b also lie in the same lattice, we have \Lambda\subset\frac{1}{N}\mathbb{Z}a+\frac{1}{N}\mathbb{Z}b. This is a full-rank lattice because a and b are linearly independent. Linnell’s theorem therefore implies that \{\pi(\lambda)f:\lambda\in\Lambda\} is linearly independent for every nonzero f\in L^{2}(\mathbb{R}) [ 4 ] . ∎ The formal Lean name of Theorem 4.1 is hrt_finite_relative_orbit . Appendix A Lean certification ledger This appendix records the exact certification status of the formal package associated with the present paper. A.1. Certified files File Status Description HRT/PhaseSpace.lean ✓ Phase space \mathbb{R}^{2} , symplectic form, integer lattices, and rational-independence infrastructure. HRT/CyclicSubspace.lean ✓ Lattice-cyclic subspace lemmas and invariance properties used in the proof of Theorem 2.1 . HRT/DenseLargeCovolume.lean ✓ Formal statement and proof of the dense large-covolume theorem. HRT/FiniteOrbit.lean ✓ Formal statement and proof of the rational-coordinate theorem. HRT/BekkaReduction.lean ✓ Formal reduction statements recorded in a separate file. HRT/CertificationReport.md ✓ Project audit report for the certified package. HRT/ImportedFacts.lean \triangleright Context object collecting the standard analytic input used by the formal proofs. A.2. Certified results Result Status Lean name Main dense large-covolume theorem ✓ hrt_dense_large_covolume

Read the complete text