The Setting
Let be the field of two elements, and let be a concrete -algebra arising from a word rewriting table of a self-simulating machine. For each integer , the Steinberg group maps into the general linear group , with image the elementary matrix group , through the representation , whose kernel is the Steinberg kernel .
The question is whether every element of becomes torsion after stabilization in rank.
Steinberg Groups and Elementary Matrices
The commutator of elements in a group is the element of .✦ Let be an associative unital ring and let be an integer. A root of rank is an ordered pair of distinct indices , with row index and column index ; the set is the set of roots of rank .✦ For each pair , the free group on has free generator .✦ The Steinberg relations in consist of the three families of relators in (1) for all : (1)✦ The Steinberg group is the group presented by the generators subject to the relators (1)✦, and the Steinberg generator is the image of in .✦
By the universal property of , for any group , a map induces a group homomorphism sending to whenever for all , commutes with whenever and , and for pairwise distinct .✦ Any ring homomorphism induces a group homomorphism sending to .✦ Likewise, for an integer , any injective map sends each root of rank to the induced root of rank and induces a Steinberg coordinate homomorphism mapping to . For an integer , the standard coordinate homomorphism is the coordinate homomorphism induced by the canonical inclusion from into .
Let be the ring of matrices over indexed by , and let the general linear group be the group of invertible matrices in . For a root and , the elementary matrix is defined by (2)✦ where is the identity matrix, has in entry and elsewhere, and the inverse of is . The elementary matrix group is the subgroup of generated by the elementary matrices for all roots and . The Steinberg representation is the group homomorphism sending to for all roots and ✦, and the Steinberg kernel is the kernel of .
Block Maps, Copy Maps, and Matrix Evaluations
The Steinberg padding homomorphism is the coordinate homomorphism induced by the inclusion from into .✦ For each root , the assignment defines a root homomorphism from the additive group of .✦ Given an integer , the canonical product bijection identifies with , sending each root and pair to the block entry root of rank .✦ As the root homomorphisms at distinct pairs commute for a fixed root , they assemble into a block root homomorphism sending each matrix to the block generator defined by (3) The Steinberg block homomorphism is the group homomorphism sending each Steinberg generator to the block generator in (3).
For each index , the Steinberg copy map is the coordinate homomorphism induced by the coordinate injection from into .✦ The diagonal copy homomorphism is the group homomorphism obtained by composing the diagonal homomorphism from to with the copy family homomorphism agreeing with on each factor .✦
Let be an -algebra, equipped with a central ring homomorphism . For , scalar evaluation is the ring homomorphism fixing and mapping to .✦ For an integer and a matrix , the numeric matrix is the matrix obtained by applying entrywise to ✦, the matrix evaluation ring homomorphism is the ring homomorphism sending to and to , and the Steinberg evaluation homomorphism is the composite group homomorphism . The cyclic shift matrix has -entry when and otherwise✦, while the truncated shift matrix has -entry when and otherwise.✦ For any ring and finite sequence in , the root word is the product .✦
The Concrete Algebra and Self-simulating Machine
For a ring and an alphabet , let be the free monoid of words over , and let be its monoid algebra, spanned by the basis monomials for . A word rewriting table over consists of a binary replacement relation on and a subset of zero words.✦ The defining relations of are all differences for pairs with together with all monomials for ; they generate a two-sided ideal ✦, and the rewriting-table algebra of over is the quotient ring .✦
For a dilation factor , a self-simulating machine hierarchy consists of a finite binary program and a threshold height such that every height satisfies the layout parameter bounds and the self-certification equivalence below.✦ From the program specialized to itself, yielding the self-specialized binary code , the hierarchy constructs at each height a verification clause schedule of capacity and a rewriting table on raw bitstrings of length with data size , whose quotient algebra over a ring is denoted . For , set and , and let binary bit-length equal for and the bit-length of for .✦ The layout bounds at height require , , , , certificate and clause counts at most , and the tape-capacity inequality (4) for the code length : (4)✦ The self-certification equivalence requires that for every , every verification task, and every Boolean assignment on data bit addresses of capacity , there exists a Boolean certificate vector on under which every clause of the height tableau holds for the assembled assignment if and only if the height codec table entry accepts the candidate words determined by the data assignment (paired with an empty right word list when the zero-task flag holds, and the true-side candidate words otherwise). Fixing the dilation factor yields a concrete self-simulating machine hierarchy of threshold height , and the concrete -algebra is the quotient algebra .
The hierarchy rewriting table acts on raw bitstrings that encode and decode finite-table cells of capacity by packing integer slots and Boolean flags. Each cell carries a syntactic skeleton (a position in a core of size , a palette of colors, or a stop position; one of five markers, idle, left, boundary, full, or right; and block metadata specifying an arity below , a zero-task flag, and a block index), a palette checker state (idle, passed, or head seeking a color), two -valued data bits, a certificate bit, and a reading head (recording side, clause index, nine-stage computation phase, and three-bit -register). A verification task pairs an arity below with a zero-task flag✦, and a verification clause schedule assigns to each task a positive clause count at most and a family of three-address clauses over data and certificate memory addresses.✦ On formatted cells (enforcing admissible markers, active-head side constraints, and endpoint skeleton bounds) and compatible adjacent cell pairs (combining structural position adjacency, internal metadata consistency, left and right marker acceptance, external boundary conditions, and palette checker compatibility), cell replacement proceeds symmetrically by local unary rules at block boundaries and local binary rules reading or undoing scheduled clauses, advancing checker heads, or swapping head-free palette cells. Zero words correspond to unformatted cells, semantic head failures or zero-task acceptance, passed palette positions, or incompatible or color-mismatched adjacent pairs; lifting cell replacement and zero conditions through the bitstring codec gives the codec table entry.
The tableau clauses are generated by a combinational not-and (NAND) circuit of two-input NAND gates over inputs, whose defining three-wire circuit clauses realize the local three-cell step of a one-tape Turing machine over a tape alphabet and state set under a one-hot binary tape label coding of bit width . In a tableau layout of parameters with , , an origin column below , certificate variable count , and clause count , addressable three-wire clauses over constant bits and memory addresses compile to verification clauses encoding initial input bits, boundary blank conditions, local circuit transitions, and origin acceptance at time . At height , the source machine layout sets , and the source machine initial row places the encoded initial symbol at origin column , followed immediately to the right by prefix-encoded program and parameter headers together with data bit references encoded via the bit symbol map, and blank symbols elsewhere.
The one-tape alphabet , state set , and transition table arise by compiling a multi-track stack program with three register tracks (including designated source and destination tracks) into one-tape commands. That program layers a self-specialization prefix wrapper over a binary packet input decoder and a universal stack machine interpreter operating on front, back, program counter, and simulated register stacks with unary ticks, separators, and stop tokens.
The Claim
Given an integer , there exists an invertible matrix satisfying (5) in : (5)✦
For any -algebra , integers and , and element , the identity (6) holds in : (6)✦
Let be an -algebra, let and be integers, and let be a nonempty finite sequence in whose polynomial entries have degree at most (with ) such that and . Then (7) holds in : (7)✦
Let be an -algebra, let and be integers, let be a nonempty finite sequence in whose polynomial entries have degree at most (with ) such that and , and suppose there exists satisfying in . Then for every , there exists an integer satisfying (8) in : (8)✦
For any -algebra , integers and , index , and element , the relation (9) holds in : (9)✦
Given an -algebra , an integer , an element , and an element satisfying in , there exist integers and satisfying (10) in : (10)✦
Over the concrete algebra , for any integer and element , there exist integers , , and satisfying (11) in : (11)✦
Theorem 1. Let be an integer and let over the concrete algebra . Then there exist integers and satisfying (12) in : (12)✦
In the paper, the result reads: let and ; then the image of in is torsion for some under the standard coordinate homomorphism. The library states the result for all in (12), expressing the standard coordinate homomorphism as .
What Is Checked
Claims (12) and (11) are stated for and elements over the concrete -algebra of the self-simulating machine hierarchy with dilation factor at threshold height , using the standard coordinate inclusions , the padding maps , and the copy maps . Claims (10) through (9) are stated over an arbitrary -algebra in (10) through (6) and (9), and over in (5), involving the padding maps and , the copy maps and , the Steinberg evaluation maps , the diagonal copy maps , the cyclic and truncated shift matrices and , root words , and polynomial evaluations and .
For the underlying constructions, for any ring and integers , the library verifies that generators satisfy for every ✦, and that both the elementary matrices in and the block generators in (for and ) satisfy the additivity, commutation, and commutator-product Steinberg relations: ✦ and ✦, commuting whenever satisfy and , with ✦ and ✦ for pairwise distinct . The library also checks that the padding index map is injective✦, that and commute in for all distinct and ✦, and that for any -algebra , scalar matrices () commute with in for every .✦ For the machine parameters, the library checks ✦, positivity of and for all , as well as positive tape coding width ✦, positive tableau clause counts✦, and positive input-packet bit width✦.