Engine · the arithmetic core

Octaves, shells, and numbers with two dimensions

Two arithmetics

Where logic flows, and where it must go around

Think of zero as an obstacle dot in a stream. The linear axis runs straight into it and stops. The exponential axis flows around it — it can pass on either side, and in Laegna it does something stranger still: it approaches zero from every direction simultaneously, so that reaching zero would require infinitely many infinities laid end to end. That is the honest sense in which a particle is less than a point. Not smaller. Differently reachable.

What each operation can reachLinear walkers land on zero. Exponential walkers close around infinity. Neither axis can do both.
+linear · zero reached · ∞ a directionwalks to zero and through it; infinity stays a direction
linear · zero reached · ∞ a directionthe mirror of addition; zero is an attainable dot
×exponential · zero never · ∞ closed aroundcloses around infinity in octaves; zero is never attained
÷exponential · zero never · ∞ closed aroundcollapses outwards; approaches zero from every direction at once
x − nΔ → 0 (attained) · x·rⁿ → 0 (never) , x·r⁻ⁿ → ∞ (in octaves)
The two axes are not two notations for one thing. They have different reachable sets.

Formally: (ℝ, +) is a group whose identity is attainable in finitely many steps from any point on a lattice, while (ℝ>0, ×) has identity 1 and treats 0 as a non-element — its closure adds 0 and ∞ as the two ends of the logarithmic line. Laegna keeps both charts and refuses to privilege either, so every quantity carries a declaration of which axis it lives on. That declaration is what dimensional analysis has been doing implicitly for two centuries; making it explicit is what lets the same engine drive a metric factor, a density of states and a musical interval.

The shells

Z, X and Y — three turns of the same dial

Here is the picture worth holding. Light speed is infinite movement inside a moment — infinite travel inside a point — so when you measure it from inside that moment it comes out as one. The space bubble of relativity is that same constant read one octave up, as an integral rather than a rate: 360° is c, 720° is the radius r of a circle whose surface is flat. Beyond 360° the moment can no longer travel; it projects downward instead, infinitesimally, as an exponent. Time and space keep their symmetry across the boundary because the boundary is the momentum repeated infinitely — a coordinate, not a speed.

Z · X · Y — one turn, three realms0°→360° is the moment out to c. 360°→720° is the bubble out to r. Below 0° lies the quantum shell.
Z-360° … 0° · Quantum realmBelow the threshold. Magnitudes here are infinitesimal in influence and unbounded in extent; they project upwards as exponents, never as sums.
X0° … 360° · Internal shell — the momentFrom rest to c. One full turn is the moment itself: infinite movement inside a point, which is why c reads as one.
Y360° … 720° · External shell — the space bubbleFrom c to the radius r whose surface is flat. The moment cannot travel here; it projects infinitesimally, as an exponent of X.

Read X as the internal Hilbert shell and Y as the external one; when only a pair of coordinates is used they map to T and R and sit parallel to Z and X, an inward direction whose composition may differ. In exponent space the merge rule changes: 0–360 is X, but 0–360 squared is Y, because the two axes are combined by exponential multiplication rather than linear sum. That is precisely why a Laegna numeral becomes two-dimensional — see the next section. The quantum shell Z, below 0°, is the same construction reflected: sub-threshold potential whose influence on the linear world arrives only as an exponent.

Z: −360°→0°  ·  X: 0°→360° ≡ 0→c  ·  Y: 360°→720° ≡ c→r
One turn is the moment; the second turn is the bubble; the reflected turn is the quantum floor.
The ladder

Acceleration is octave −1, and it must be held

Acceleration sits one octave below velocity: it is differential 1, the derivative rung. Velocity is octave 0 — light speed, differential 0 and integral 0 at once, which is why c reads as the unit of the whole ladder. Position is octave +1, the integral rung, the place where growth still has room; take that growth to infinity and you get c back, now as the radius of the bubble. Even at the derivative rung, c is the largest quantity that could be added to light: past it, light speed would be the thing obeying, not the thing limiting.

Acceleration is a tension, not a pushβ = tanh(aτ). Rapidity adds without limit; velocity saturates at the pole. Release the tension and growth stops.
-1acceleration · differential 1A positive tension that must be kept for velocity to grow at all. Even here c is the largest quantity that can be added to light.
0velocity · c · differential 0 · integral 0The pole read as unit. Infinite movement inside a moment, so it measures as one and cannot be exceeded.
+1position · r · integral 1Where growth still has room. Taken to infinity it equals c again — the radius of the space bubble.

The relativistic rocket equation β = tanh(aτ/c) makes the ladder explicit: rapidity φ = aτ/c is additive and unbounded, β is bounded and saturating, and the two are related by the same compression map used everywhere on this site (tanh is compress() in exponential coordinates). Proper time along constant proper acceleration, τ = (1/a) asinh(at), is its integral — the exponential clock. See the relativity page for the full treatment and the critical numbers.

Notation

A number is a matrix of quarter turns

Four glyphs — I, O, A, E — are the four quarter turns, so a digit is a direction and a row is a path. Because 4 = 2², every digit is exactly two octaves and every four-digit row is exactly eight: a row is a block of 256, the largest number most people still hold as a single thought. Stack rows and you are no longer adding, you are exponentiating, which is how the notation itself encodes the X→Y merge. The number system stays harmonically synchronised with infinity: reaching further in octaves never invalidates the smaller numbers, because the axes do not yield.

A number as a matrix of turnsWithin a row: linear addition of quarter turns. Between rows: exponential blocks of 4⁴ = 256.
× 2560
AOEO
= 118
× 2561
OAEA
= 185
× 2562
IAEO
= 120
value = 7,911,798  ·  octaves = 22.92  ·  every digit is exactly two octaves, every row exactly 8
n = Σr ( Σc drc·4c ) · (4cols)r
Rows sum linearly; the stack of rows composes exponentially. Notation follows geometry.

The construction is a mixed-radix positional system in which the outer radix is itself a power of the inner one, so the map to binary is exact at every level (log₂4 = 2, no rounding drift) and the two-dimensional layout carries the additive and multiplicative gradings separately rather than flattening them. Practical consequence: fixed-point arithmetic in this notation keeps the octave count of a quantity as metadata for free, which is the piece that ordinary floating point throws away and then reconstructs by exponent bookkeeping. Whether this yields a genuinely cheaper representation for scale-heavy simulation is an open, testable, and entirely ordinary piece of work.