Native Alternating Switching Polygons
Abstract
The literal native T and E switching words close, and their actual prefix traces have no repeated vertex before return.
Configurations are actual permutations of Fin(m+1), and right multiplication acts on positions. The reversals a, b and c have lengths m+1, m and m-1. Put B=bc. The halfword is the list [a,c] followed by m-2 copies of [b,c], flattened. The T word repeats this halfword twice; the complete E word consists of m copies of [b,c]. All repetition here preserves the literal generator list. The trace is scanl of right multiplication, so it records every actual prefix product.
Theorem 1.1 (The Exact Halfword Endpoint).
Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/PrefixReversal/NativeSwitches.native_t_halfword_swap (✓ std3). ∎
Source. Repository-derived.
Commentary.
For every m at least three, the halfword product is precisely the swap of positions zero and m. The proof computes every power of bc on each native position, and then applies a and c. It does not assume an endpoint matching or a finite enumeration.
Theorem 1.2 (The Complete T Polygon Is Simple).
Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/PrefixReversal/NativeSwitches.native_t_simple_trace (✓ std3). ∎
Source. Repository-derived.
Commentary.
The T product is identity and its prefix trace with the final return removed has no repeats. Writing tau for the endpoint swap, the trace before return is [1,a], the paired interior ports, [tau,tau*a], and the tau-left-translates of the same interior ports. The last-label-zero group has native coordinates (m-2,false), (m-1,true), and both orientations at all indices below m-2. The existing coordinateEquiv proves they are distinct. Left multiplication by tau preserves distinctness and moves their last label to m, which separates the two groups.
Theorem 1.3 (Every Complete E Supplier Is Simple).
Proof. Machine-checked in Lean as D5/S3/Combinatorics/Graph/PrefixReversal/NativeSwitches.native_e_simple_trace (✓ std3). ∎
Source. Repository-derived.
Commentary.
For every starting configuration z, the complete E word has product identity. Its actual trace is the ordered flattened list [zB^j,zB^j*b], for j in Fin(m), followed by z. Each pair has the two different native reflection coordinates at one gap; different gaps have different coordinates. The existing coordinateEquiv therefore proves no repeat before the return.
These results establish actual native switching polygons. They do not establish that a toggled factor is connected, that each cut belongs to a different complete native circle, or that the three-generator Cayley graph has a Hamilton cycle. Those support and splicing obligations must be proved separately.
References
- Truth anchor:
D5/S3/Combinatorics/Graph/PrefixReversal/NativeSwitches.native_e_simple_trace - Truth anchor:
D5/S3/Combinatorics/Graph/PrefixReversal/NativeSwitches.native_t_halfword_swap - Truth anchor:
D5/S3/Combinatorics/Graph/PrefixReversal/NativeSwitches.native_t_simple_trace - Dependency: D5/S3/Combinatorics/Graph/PrefixReversalZeroStarComplement