How infinite values preserve posture through symbolic atanh/tanh fusion.
Every value in SSM-Infinity — including (+∞) and (-∞) — carries a bounded alignment lane:
a ∈ (-1, +1)
The alignment lane is not “probability,” not “weight,” and not “uncertainty.”
It is a symbolic posture representing how the value enters the arithmetic.
Through this lane, even infinite values can take part in lawful, reproducible merging.
SSM-Infinity uses the same alignment kernel as the rest of Shunyaya Symbolic Mathematics:
a_c := clamp(a, -1+eps, +1-eps)
u := atanh(a_c)
U := U1 + U2 # merging posture in hyperbolic space
a_out := tanh(U)
This hyperbolic method ensures three key properties:
- Determinism — no randomness, no branching ambiguity.
- Reversibility — alignment behaviour is mathematically reversible.
- Boundedness — output always remains inside (-1, +1).
1. Why atanh/tanh? Symbolic Geometry of Stability
Classical merging (average of weights) does not work for infinities.
It collapses easily or becomes unstable.
Hyperbolic merging is chosen because:
atanh()linearizes the bounded intervaltanh()re-bounds the final posture- It is order-invariant
- It is associative (within safe precision)
- It is stable for extreme inputs, including infinite-class values
This gives SSM-Infinity its signature behaviour:
(+∞, a1) + (+∞, a2) → (+∞, a_out)
(+∞, a1) + (-∞, a2) → zero-class # collapse into stable symmetric class
2. Alignment Lane in Infinite Operations
Addition (same direction)
a_out = tanh(atanh(a1) + atanh(a2))
Addition (opposite direction)
The sign conflict forces a zero-class collapse, discarding alignment:
(+∞) + (-∞) → zero-class
Multiplication & Division
The alignment lane simply travels with the sign:
(+∞, a) * (-5) → (-∞, a)
(12) / (-∞, a) → zero-class
Exponentiation
The lane remains unchanged except for zero-class outcomes:
(+∞, a) ** (-k) → zero-class
In SSM-Infinity, alignment always respects the symbolic class rules.
3. Relevant Code (Hyperbolic Merge)
Below is the exact merging logic from the engine:
# hyperbolic merge (excerpt from ssm_infinity_core)
def merge_align(a1, a2):
# clamp inputs
a1_c = max(min(a1, 1-1e-12), -1+1e-12)
a2_c = max(min(a2, 1-1e-12), -1+1e-12)
# map into hyperbolic posture space
u1 = math.atanh(a1_c)
u2 = math.atanh(a2_c)
# combine and re-bound
return math.tanh(u1 + u2)
This is the same kernel used across the entire Shunyaya ecosystem —
SSMDE, SSMNET, SSMEQ, SSMClock, and now SSM-Infinity.
4. Why It Matters
Thanks to hyperbolic alignment merging:
✔ Infinite arithmetic becomes predictable
✔ Symbolic class boundaries stay clean
✔ Extreme values never destabilize the computation
✔ The entire system becomes reproducible and testable (22/22 tests passed)
In other words:
SSM-Infinity does for ∞ what SSM-Zero did for 0 — it makes it lawful.
Page Navigation
Previous: SSM-Infinity — Infinite Arithmetic & Deterministic Outcomes
Next: SSM-Infinity — Core Engine (Full Source Code)
Directory of Pages
SSM-Infinity – Table of Contents
Disclaimer
Shunyaya Symbolic Mathematical Infinity (SSM-Infinity) is a symbolic research framework — not a predictive model or numerical estimator.