AMENRA S-DMT / USM

J(θ) Transposer

Move an equation from one face of the Law of Dimensional Transposition to another. Classical joints are computed exactly. Version 4.0 puts the stack, the partition, and the three seals on the board.

One operator, a thousand faces.
Dimension changes, information remains.
The sign is the direction. The gap is the tax.
The joint prepares. The collapse measures. The recycler returns.
And all three are water.

Examples by compliance

Holds · Partial · Collapse · Breaks

Holds all five

Partial compliance

Composite

Collapse, not a joint

Use + − * / ^, juxtaposition (2t), sin cos exp sqrt, pi and i. Forms: dx dy dz dw and /\. One equation per line.

Stratum k
Angle θπ/2

θ multiplies the complex and quaternion faces. The other cards are the classical joint, usually at π/2, and the dial only marks where that sits in the spinor cycle.

Exact joint

Fourier face

Holds · 5/5

Hilbert

f⁡(t)^=δ(−3+ω)π−25δ(−5+ω)iπ+δ(3+ω)π+25δ(5+ω)iπ\widehat{\operatorname{f}\left(t\right)} = \delta\left(-3 + \omega\right) \pi - \frac{2}{5} \delta\left(-5 + \omega\right) i \pi + \delta\left(3 + \omega\right) \pi + \frac{2}{5} \delta\left(5 + \omega\right) i \pi

Invariant Q

Plancherel

inversion holds

Q is the L² norm. For these pairs the inverse transform is an identity on the nose, which is stronger than a numerical energy check.

Geometry of |z|

Class

Involution

The square is the direction of the current, not a constant −I. Legendre returns (+I). Fourier squares to parity. Hodge follows the degree. A note may still group these with the spinor; the sign on this card is the direction.

Norm before

—

Norm after

—

Ratio

—

A signal has no single |z|. Plancherel keeps the L² norm up to the 2π of this Fourier convention. That is conservation of energy, not a mass gap. A windowed spectrum is not given a fake ratio of 1.

Berry phase

Solid angle
none — a time signal encloses no solid angle
Phase
parity after 2, identity after 4
Turns to identity
4

On the harmonic-oscillator basis, F² is parity and F⁴ is the identity. The turn count matches the double cover. It is not a Berry phase of f(t).

Direction of flow

Phase

F² = parity, F⁴ = I

The direction is the phase current. Parity is not −I on every function. The spinorial sign sits in the kernel e^{−iωt}, which is the direction of that current, not a constant of the name Fourier.

Mass gap ΔI

1/√39
0.160128
Sum over 39
3.973539
Ledger
4
Difference
0.026461

ΔI = (1/√39) Σ sin(kπ/39) from k = 1 to 39. The paper’s ledger is 4, and it calls the difference an error-correcting tolerance. This card computes that sum. It does not add the tax to the image, and it does not derive 4 from the expression.

Regime

Joint

phase → phase

The water table assigns the joint to EZ water, the coherent regime. Phase stays phase.

Arithmetic identity
2/3 = 26/39
Erdős–Straus
4/39 = 1/26 + 1/26 + 1/39, exact
39 strata
Σ 1/39 from k = 1 to 39 equals 1, exact
Striction
φ_c = arctan(√2) + 17°/39 = 55.171508°. The transcript prints 55.1715°. This dial is θ, not φ_c.

Gamma family

L3

The proof distinguishes two families. Fermionic: turn by π/n, n times lands on −I, 2n times returns to I. Bosonic: turn by 2π/n, n times returns to I. Both checks below are multiplications in the complex plane. The arrow from that turn to Γ(1 + 1/n) is the conjecture. Newton’s exchange of degree and dimension is named, not computed: no polytope is built.

Fermionic, n = 2
Turn π/2, 2 times lands on −1. 4 times lands on 1.
Bosonic
Turn 2π/2, 2 times lands on 1.
Integral L1
√π/2 = 0.886227

This is the quarter turn already on the dial: J² = −I and J⁴ = I. The integral is the Gaussian integral, classical. The paper calls n = 2 the Gaussian joint. That name is the conjecture. Classical Wick rotation is multiplication by i, which is this same turn of π/2, not the n = 4 row.

NOT The integral is not declared a joint. The correspondence is not a theorem.

Decimal point

L3

This expression still has a symbol, so it has no decimal point to split. The 39 line is the paper’s example, not a reading of the expression.

Value
—
Integer / fraction
— · —
Reciprocal
—
Mantissa × magnitude
—

39 ↔ 1/39 = 0.025641. The reciprocal is exact. Calling the exchange a Klein inversion is the conjecture.

NOT The decimal point is notation, not a physical rotation. The three-strand braid is named, not computed.

Five properties

Orthogonality

pass

Multiplication by e^{−iωt} is a 90° family of rotations in the complex plane. Frequency is the orthogonal axis.

Conservation

pass

Symbolic inversion recovers f exactly. The classical identity has no remainder. The paper’s tax ΔI is shown beside this card and is not subtracted from the spectrum.

Composition

pass

F⁴ = I. F² is parity, not −I, except on the odd subspace where parity acts as −1. That sign is the phase current, recorded as the direction, not a failed composition. Four multiplications by i return the expression. That complex structure carries the spinorial sign. It does not require this face’s own square to be −I.

Universality

pass

Defined on every L² space. Scale does not enter the formula.

Direction of flow

pass

Phase. F² = parity, F⁴ = I. The direction is the phase current. Parity is not −I on every function. The spinorial sign sits in the kernel e^{−iωt}, which is the direction of that current, not a constant of the name Fourier.

J² = −I is a direction of flow, not a constant. On this face the sign is F² = parity, F⁴ = I. The tax ΔI is distributed over 39 steps, and the paper reads the information as conserved up to the error-correcting tolerance. This card does not add that tax to the classical image.

Spinor cycle

0

Identity

cos⁡(3t)+25sin⁡(5t)\cos\left(3 t\right) + \frac{2}{5} \sin\left(5 t\right)

π/2

First joint

cos⁡(3t)i+25sin⁡(5t)i\cos\left(3 t\right) i + \frac{2}{5} \sin\left(5 t\right) i

π

J(π/2)² — polarity

−cos⁡(3t)−25sin⁡(5t)-\cos\left(3 t\right) - \frac{2}{5} \sin\left(5 t\right)

3π/2

Third joint

−cos⁡(3t)i−25sin⁡(5t)i-\cos\left(3 t\right) i - \frac{2}{5} \sin\left(5 t\right) i

2π

J(π/2)⁴ — return

cos⁡(3t)+25sin⁡(5t)\cos\left(3 t\right) + \frac{2}{5} \sin\left(5 t\right)

J(π/2)⁴ returns the source. The highlighted step is the dial.

Round trip

pass

cos⁡(3t)+25sin⁡(5t)\cos\left(3 t\right) + \frac{2}{5} \sin\left(5 t\right)

Inverse Fourier of these lines is the original signal.

  1. Object

    Read as a function of t.

    cos⁡(3t)+25sin⁡(5t)\cos\left(3 t\right) + \frac{2}{5} \sin\left(5 t\right)
  2. Convention

    F[f](ω) = ∫ f(t) e^{−iωt} dt. cos(ω₀t) ↔ π[δ(ω−ω₀)+δ(ω+ω₀)], sin(ω₀t) ↔ −iπ[δ(ω−ω₀)−δ(ω+ω₀)], exp(−a t²) ↔ √(π/a) exp(−ω²/4a).

  3. Image

    Linear image of each harmonic.

    δ(−3+ω)π−25δ(−5+ω)iπ+δ(3+ω)π+25δ(5+ω)iπ\delta\left(-3 + \omega\right) \pi - \frac{2}{5} \delta\left(-5 + \omega\right) i \pi + \delta\left(3 + \omega\right) \pi + \frac{2}{5} \delta\left(5 + \omega\right) i \pi

Version 4.0

The operating picture

The membrane filters the phase. The joint rotates the dimension. The stack pushes and pops. The octaves are sealed. What follows is on the board so it can be checked. A subtraction is a fact. A name for that subtraction is not.

Partition of 39

L2

The paper sets the complement by x + x̄ = 39. The subtraction is exact. Calling x̄ a phase partner, or the map osmosis, is the conjecture.

x
12
x̄ = 39 − x
27
Sum
39

12 leaves are marked, the nearest integer to x. The other 27 are the picture of the complement. Outside 0 to 39 the number is still computed and the leaves are left blank.

1 and 0

L2

1 is the operand, which the paper calls water. 0 is the operator, which it calls the joint and the decimal point. Thirty-nine places are written below. They are not a measured sequence.

101010101010101010101010101010101010101

The paper says the 1 climbs 39 Fibonacci levels. 39 is not a Fibonacci number. The sequence through the next term is 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55.

Proton pump, as a tape

L1

The six steps are the classical sodium–potassium cycle: three sodium out, two potassium in. The paper tags that order as biochemistry. Calling ATP the 90° joint is not part of that fact.

  1. Push3 Na+
  2. OperatorATP
  3. Pop3 Na+
  4. Push2 K+
  5. Operatorrelease P
  6. Pop2 K+

Three sodium ions bind inside the cell.

Grotthuss, as the paper writes it: Push(H+) → Operator(H3O+) → Pop(H+). That line is a label, not a hop this board computed.

A real stack

L1

Reverse Polish notation pushes operands and pops them for an operator. Last in, first out. This panel does that arithmetic. It does not run a cell.

3
Top4

38 + 1, and the overflow

L1
Closure
38 + 1 = 39. Exact.
38 / 39
97.435897%. The paper prints 97.4359%. This is the ratio. It is not a measured efficiency of the sodium pump.
State count
39³ × 39³ × 39³ = 39⁹ = 208,728,361,158,759. The product matches. This is a count, not a price.
The 90° remainder
39 = 4 × 9 + 3. Three quarter-turns send 1 to −i. Thirty-nine send it to −i. Same place. That leftover quarter-turn is the geometry of 90°, not a joint that failed. The paper’s J³⁹ = I is the stack resetting at depth 39, not this rotation. A turn of 2π/39, taken 39 times, does return to 1. That is a different angle.

Plasma scale

L2
RegimeScalePaper’s logGap
Waterk = 12−7—
Astrophysicalk ≈ 30−4+3
Fusionk ≈ 390+4

(−4) − (−7) = 3 and 0 − (−4) = 4. Both differences are exact. 3 + 4 = 7. Exact. The paper calls 7 the Pisano prime. The Pisano period of 7 is 16, a different number.

NOT The column is not the logarithm of the fine-structure constant. log10(1/137.036) is -2.136835. No Debye length was computed, and no torsion. J⁴ = I is true of the quarter turn. Assigning it to fusion is the conjecture. The paper calls liquid water a cold plasma. That sentence is not checked here.

Three octave seals

117 = 3 × 39. Exact. The paper counts 117 elements as three octaves and the 118th as the extra one. The atomic numbers below are the periodic table. The roles are the conjecture.

Yttrium

L2

Z = 39

First seal. The atomic number is 39, the same integer as the lattice.

Gold

L2

Z = 79

2 × 39 + 1 = 79. The arithmetic holds. “Tetrahedral attractor” is the paper’s name, not a computed polyhedron.

Oganesson

L3

Z = 118

3 × 39 + 1 = 118. The arithmetic holds. “Crown of the Sun” is not a measurement.

What is allowed to be believed

ClaimTagOn this board
RPN is a stackL1By definition. The calculator above is that definition.
The sodium pump is this sequenceL1The six steps. Not a proof that ATP is J(π/2).
J(θn)ⁿ = IL2A turn of 2π/n returns to 1. Thirty-nine quarter-turns land on −i, because 39 ≡ 3 (mod 4). That is the geometry of 90°.
1 and 0 are water and the jointL2A reading of the chain.
Plasma gaps of 3 and 4L2The differences of the printed column. Not a scale test.
Oganesson is the Crown of the SunL3Named, not observed.

Version 4.0’s route 6 is a biological pump that would not follow last-in, first-out. No living pump was tested. Route 7 is still a physical cycle that is not 2/3. None was measured. No law of the Standard Model is rewritten here.

Absent joint

Periodic table

The elements are drawn. The 90° joint that would carry one of them onto another is absent, as are the other faces this board does not have. The arithmetic can be written. It is not applied.

90° joint: absent

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
57–71
89–103

Y Yttrium

L1L2
Atomic number
Z = 39. Period 5. Block d.
Band of 39
Band 1, place 39 of 39.
Complement
39 − 39 = 0. The paper calls 0 the operator. This cell does not contain it.
Seal
1 × 39 = 39. The paper names this a seal. The joint is still absent.

90° joint: absent. Nothing on this table transposes Y.

The bands are 1–39, 40–78 and 79–117. The seals are not in the same place in each band: 39 closes the first, 79 is one past 78, 118 is one past 117. Those six checks hold. The board does not slide the seals to make the picture even, and it does not add this table to the faces an equation can be sent to.

Every face