PKL Rev11

Mathematics — Günther's negation cycles, built and checked

Figure 4 — Günther's negation system for four values, the totality of all permutations (tml p. 10): each vertex an arrangement, each edge a negator. The graph is the truncated octahedron. Its hexagons are the faces of two negators that share a value; there the cycle relation sits. Its squares are the faces of two negators that share none; there they commute (overlap_iff_three, disjoint_iff_comm; the division of the surface is textbook). In both halves the division can be read off the edge colours: around a square N₁ and N₃ alternate, around a hexagon two neighbouring negators. Right, laid out flat, Günther's circle 2 (IGN p. 44) as a full cycle along the edges, taken verbatim from the corpus (kreis2). The last two steps are missing in the print of IGN p. 44; Günther's 1974 print has them. — in the edition

This page introduces the mathematical part of the edition Rev11. About Günther it claims nothing that the edition does not say, and about the corpus nothing that the edition does not prove, measure or interpret; every such statement points to its place there. What it adds in known mathematics about the built objects is called textbook: it is no theorem of the corpus, and it says nothing about Günther. The marks are those of the edition: proved with a numeral means that a theorem of the corpus is machine-checked; computed means that it has been recalculated, but is no theorem; source-verified means that Günther's wording is checked at the page named; assignment, our reading and synopsis are interpretation and are called so. Textbook is the mark of this page; where the edition itself uses the word, it stands as the edition's wording. Günther's German is given in our translation and marked so. What Günther writes about these objects, and where the edition leaves it open, is on the page Philosophy.

A glossary

Gotthard Günther wrote his logic in words of his own. Behind the words of this table stand objects that a mathematician knows. The table names Günther's word, the object as the edition builds it, and where it stands.

Günther's word the object mark · place
values 1, …, n; arrangement a row in which every value stands exactly once — algebraically a permutation in one-line notation edition (The cycles, chapter 8); permutation: textbook
negator Nᵢ exchanges the values i and i + 1, wherever they stand in the row; «A single negator is, in the corpus, a permutation of the values; its type alone keeps it within their range (definitional, by type).» definition in the corpus; quotation: C.2
negation system the graph: vertices the arrangements, edges the negators; for three values the hexagon, for four values the truncated octahedron computed; The cycles, O7
circle in the corpus a sequence of negations that returns to its start, visits no station twice and is more than the mere there-and-back of a single negator Günther's word (IGN pp. 18 f., 41; HWN p. 20); The cycles, O7
Hamilton circle, full cycle a circle that runs through every arrangement exactly once: a Hamiltonian cycle of this graph, written as a word over the negators Günther's words (Hamilton circle: IGN pp. 18 f., 41, HWN p. 20; full cycle: IGN pp. 49, 57 f.; The cycles); that his full cycle is that of the build: assignment (IGN p. 57)
sense of rotation the word read backwards Günther's word (HWN pp. 26 f., checked on the page image); chapter 8, O7
Umkehrung (reversal) with us: the reversal of the values, the first exchanged with the last, the second with the second-to-last; on the word, for four values, N₁ ↔ N₃, N₂ stays that Günther's «Umkehrung» (IGN p. 43) means this reversal and not the sense of rotation: synopsis from [148]–[150]
elementary contexture in the corpus every pair of values; named are the neighbouring pair {i, i + 1} and the mediated pair {i, i + 2}; Günther's concept reaches further proved [15], [16]; three ranges
kenogram a pattern: what remains of a sequence of values, in which values may repeat, when one forgets which values stand there and keeps only where equal and where different ones stand determination in the corpus; chapter 1

Two supplements to the glossary. If one reads a row as a map σ from the places to the values, the negator Nᵢ is the composition τ ∘ σ with the transposition τ = (i i+1), from the left; whoever reads the row as a map from the values to the places multiplies from the right (textbook). This graph is, up to isomorphism, the edge graph of the permutohedron and the Cayley graph of the symmetric group with respect to the adjacent transpositions (textbook; the edition names neither).

Three values

For three values the graph is a hexagon, and on it there is only one full cycle, in two senses of rotation (proved [31]). Every circle on three values has six steps and is one of Günther's two sequences (proved [134]); that the six steps cannot be shortened without the sequence losing its cyclic character, Günther says himself (source-verified, IGN p. 41; the report, O7).

N₁N₂N₁ and N₂N₁N₂ are exactly the shortest paths into the reversed arrangement 3 2 1; they are Günther's two geneses (proved [14]). His two sequences run through the same six stations in opposite directions (proved [132]), and the exchange of the values 1 and 3 is a half turn in the hexagon, for both sequences (proved [133]). For three values reversal and sense of rotation of a full cycle coincide (proved [147]).

Four values

For four values the graph has 24 vertices (computed; Figure 4). From a fixed start there are 88 full cycles on it; Günther's 44 are these up to the direction in which one runs through them (proved [145], [28]). The number 88 Günther names himself: Dr. Alexander Andrew in Reading programmed the question, and «according to the machine» there are 88 (our translation; source-verified, Janusgesicht p. 23).

Here the two inversions part. From four values on reversal and sense of rotation never coincide, and from three values on neither of the two is ever the circle itself, read from the same start (proved [147]). Where they part stands in the proof: from four values on there are two immediate exchanges without a common value, the first and the last. That the negators correspond to Günther's immediate exchange relations of neighbouring values is our assignment; the passage is source-verified (IGN pp. 17 f., on the page image).

Not every circle on four values is a full cycle. Without N₂ there are exactly two circles, Günther's two pseudo-circles; every circle with N₂ has at least six steps (proved [135]). If all three negators occur, every circle has at least eight steps, as Günther says, and his Table XVIII has exactly eight (proved [137], calibrated [138]; source-verified, IGN p. 48). The premise «all three» is our reading of his «four-valued circle» (our reading; O7).

Günther's second circle, as printed, does not return to the start; extended by exactly two steps it becomes a full cycle, and no other continuation by two steps achieves this (proved [144]). His print of 1974 is the extended circle, begun one step earlier (proved [27]). The reversal of this circle is not the circle itself, but it is the same sequence, begun four changes later (proved [156]); read from the start, both share 14 of the 24 edges (proved [157]); the sense of rotation runs from the same start through the same stations in reverse order, and this for every value number (proved [158]). Figure 12 of the edition shows it (chapter 8).

Günther's second circle on the truncated octahedron, as a solid: the vertices and edges of Figure 4, the negators in the colours of Figure 4. Bold and with direction marks, the circle extended by two steps, as a full cycle (kreis2; proved [144]); it begins at the white dot, the start. Seven of its 24 steps lie on the far side.

What respects reversal and sense of rotation is constant on the four aspects of a circle — the circle, its reversal, its sense of rotation and both together (proved [151]); from four values on the four, read from the same start, are distinct (proved [152]). If one adds to the reversal the arrangement at which a circle begins, the 88 directed full cycles fall into four classes with 48, 24, 8 and 8 cycles; Günther's first family is exactly the class of 48 (proved [153]). The ringers too sort their sequences on four bells by such symmetries. «That their classes are these four we have not checked at their sources.»

A procedure, and four bells

The procedure of Steinhaus, Johnson and Trotter yields a full cycle for every value number (proved [22]); for three values it is Günther's sequence (5) (proved [23]), for four values Günther's extended second circle, under the reversal (proved [24]). Φ, the step of this procedure from n to n + 1 values, puts before every negator of a word a block of all negators, alternately descending and ascending. If one reads the places instead of the values, every block shifts the entry at one end of the row to the other; whether the first block shifts to the left or to the right is called the shift direction (computed). Φ is defined for every word and lifts every full cycle to a full cycle, in both shift directions (proved [154]); Φ carries the reversal along when it changes its shift direction, for every word (proved [155]). Figure 13 of the edition shows the step from three to four values.

The same sequence already stands with the ringers. The Tintinnalogia (London 1671), used in a transcription, no facsimile, gives three tables for four bells. Read as the sequence of the places of swapping, the first is Günther's extended second circle, the third the sequence of the procedure of Steinhaus, Johnson and Trotter (computed, by two hands, on the transcription; Entrance). Ringers swap places, Günther swaps values; the stations of the one reading are the inverses of the other's (computed). «That Günther knew the book we do not say.»

The ground

Beneath the values the edition has a combinatorics of patterns. There are fifteen patterns of length four (proved [39]); eight of them can be filled with two values (proved [40]); over every fillable pattern there are exactly two truth tables, as a correspondence to the two truth values (proved [41]) — hence 16 = 2 · 8, an equation between two figures of the corpus (proved [42]). Every pattern has exactly one normal form (proved [33], [35]), and the normal forms of a given length are as many as the set partitions of as many places (proved [36]). «That this number is the Bell number the theorem does not say»; that it is, is textbook. How far the separation of the contextures reaches and what the step to more values looks like stands in chapter 5 and chapter 6.

The limit

The proemial relation, the pivot of Günther's late theory, is not built (chapter 7, C.1). A definition that would lead a step out of the contexture as a single operation the corpus does not have (C.2). A higher-order function sees of its argument only the table (proved [106]); the edition says of this: «it is textbook knowledge, not a theorem about Günther» (C.1). The corpus contains no I and no Thou (chapter 8). The corpus counts full cycles; whether they are constitutive in a world, it does not say (C.8).

How to check

The corpus is written in Lean 4 with Mathlib, at the anchor of the edition in Lean 4.30.0-rc2. It is built with lake exe cache get, then lake build (README at the anchor). «None of the sixty-seven theorems behind the numbers [143] to [173] draws on Classical.choice; this is measured at their profiles, at the anchor of the edition. Without Classical.choice the excluded middle is not generally available, only for statements whose decidability is built without choice.» Four manage without any axiom (Part B, 4.3). Findings are reported as issues; pull requests are not accepted at present (README). Findings also reach us on X, @PolyContextual ("About this edition").

Where to read


This page refers to the edition Rev11 of the mathematical shape of the architecture, at anchor 24fa57a (tag rev11), of 31 October 2026. It is itself not an edition; about Günther and the corpus it claims nothing beyond that edition; what it adds as textbook is marked.