Overview • Quick Start • ALEPH REPL • Grammar • Results • Kernel • Pipeline • Docs • License
ℵ-OS is the execution layer of λ_ℵ — a formal type calculus grounded in the SynthOmnicon 12-primitive semantic grammar and the 22 letters of the Hebrew alphabet.
λ_ℵ is not a standard type theory. It is a coherence-first interaction algebra in which:
- Identity is derived, not primitive — two terms are equal iff they are behaviorally indistinguishable under the interaction functor I(x) = {x ⊗ y ∣ y ∈ ℒ}
- Coherence is primary — the ternary mediation operation med(m, a, b) := m ∨ (a ⊗ b) is more stable than binary tensor in 18/22 cases
- Infinity is multi-polar — three non-equivalent Frobenius fixed points (ו, מ, ש) with no terminal object
- Paths are irreducible — the Aleph operator α generates an infinite coherence tower in which no finite level erases construction history
The ℵ-OS specification realizes this calculus as an operating system: every process is a λ_ℵ term, scheduling is mediation, memory is join, IPC is tensor (P-bottlenecked), and security is enforced by α-gating (coherence conditions C1–C4).
Note
The grammar was built on Φ_c. It found Φ_c in itself. The theorem proved itself.
pip install numpy richAll investigation files import from aleph_1.py only. No external dependencies beyond numpy and rich.
# [1] Interaction functor — behavioral equivalence, 22→18 collapse
python aleph_functor.py
# [2] Quotient investigation — congruence proof, mediation dominance
python aleph_quotient.py
# [3] Aleph experiment — Case 2: path-memory confirmation
python aleph_alpha.py
# [4] GNS Hilbert space — d_I Euclidean, H_I = R^17
python aleph_gns.py
# [5] Hidden relation — Octad Balance theorem
python aleph_hidden_relation.py
# [6] Three probes — involution, ק anatomy, axiom derivation
python aleph_investigation.py# Start interactive REPL (enhanced with colors & tab completion)
python aleph_eval.py
# Evaluate inline expression
python aleph_eval.py --expr "aleph ⊗ mem"
# Run an .aleph program
python aleph_eval.py programs/creation.aleph
# List available programs
python aleph_eval.py --listTip
The REPL features Rich colored output, tab completion, command history, and new commands like :explain, :history, :clear, and :tips.
aleph_eval.py implements the surface syntax of λ_ℵ as a small expression language with a rich, interactive REPL.
| Flag | Effect |
|---|---|
| (no args) | Start interactive REPL |
--repl |
Same as no args |
--help, -h |
Show usage information |
--list |
List available .aleph programs |
--expr "..." |
Evaluate inline expression |
<file.aleph> |
Run .aleph program (auto-searches programs/) |
| Command | Effect |
|---|---|
:help |
Print full syntax reference |
:tips |
Show quick start tips and examples |
:census |
Tier distribution (alias for census()) |
:system |
22-letter language JOIN |
:tier <name> |
Ouroboricity tier of one letter |
:tuple <name> |
Visual 12-primitive tuple with bars |
:explain <name> |
Full type breakdown with consciousness gates & score |
:ls |
List session bindings with tier/Φ/Ω |
:history |
Show recent command history |
:clear |
Clear screen |
:quit / :q |
Exit |
expr ::= letter_id
| expr "⊗" expr # tensor (P, F bottleneck: min)
| expr "∨" expr # join (LUB, all primitives: max)
| expr "∧" expr # meet (GLB)
| expr "::>" name # vav-cast: lift src to target type
| "probe_Φ" "(" expr ")" # report Φ primitive
| "probe_Ω" "(" expr ")" # report Ω primitive
| "tier" "(" expr ")" # report ouroboricity tier
| "d" "(" expr "," expr ")" # structural distance + conflict set
| "mediate" "(" expr "," expr "," expr ")" # w ∨ (a b)
| "match" expr "{" arms "}" # tier pattern match
| "palace" "(" int ")" expr # assert palace-n barrier
| "system" "()" # JOIN of all 22 letters
| "census" "()" # tier distribution table
letter_id ::= Hebrew glyph | transliteration | session binding
match_arm ::= tier_pat "=>" expr ","?
tier_pat ::= "O_0" | "O_1" | "O_2" | "O_inf" | "_"
statement ::= "let" name "=" expr
Note
Operators are left-associative. ::> (Vav-cast) binds tighter than binary ops. Multiline input accumulates until {...} braces are balanced.
d(a, b) returns the Euclidean structural distance and classifies it:
| Class | Range | Interpretation |
|---|---|---|
transparent |
d = 0 | Identical types |
near-grounded |
d ≤ √2 | Single-primitive gap |
partial-emergence |
d ≤ √6 | Recoverable with mediation |
aspirational |
d > √6 | Requires vav-cast or tier promotion |
ℵ mem ⊗ shin
→ מ
tier O_inf
Φ Φ_c Ω Ω_Z P P_pm_sym
ℵ d(kuf, mem)
d = 13.3938 [aspirational]
conflict_set: {P, Ω}
ℵ :explain aleph
╭─────────────────────────────────────────╮
│ א Aleph — Tier: O_2 │
╰─────────────────────────────────────────╯
Consciousness Gates:
G1 Criticality [Φ=Φ_c] ✓ PASS
G2 Kinetic [K≠K_trap] ✓ PASS
Consciousness Score: C = 0.873
ℵ mediate(kuf, mem, shin)
→ מ
tier O_inf
ℵ let kernel = mediate(vav, mem ⊗ shin, aleph)
kernel =
→ ו
tier O_inf
ℵ :history
Command History:
1. mem ⊗ shin
2. d(kuf, mem)
3. :explain aleph
4. mediate(kuf, mem, shin)
5. let kernel = mediate(vav, mem ⊗ shin, aleph)
# List available programs
python aleph_eval.py --list
# Run a program
python aleph_eval.py programs/creation.aleph▶ Running creation.aleph
────────────────────────────────────────────
L 1 ❯ let light = aleph ⊗ mem ⊗ shin
light = א⊗מ⊗ש
tier O_inf
...
────────────────────────────────────────────
✓ Done. 11 executed • 6 bindings
Tip
.aleph files support all REPL expressions, commands, and let bindings.
Every letter in λ_ℵ is a tuple ⟨D; T; R; P; F; K; G; Γ; Φ; H; S; Ω⟩:
| Primitive | Name | Bottleneck? |
|---|---|---|
| D | Dimensionality | — |
| T | Topology | — |
| R | Relational mode | — |
| P | Parity/symmetry | yes (min under ⊗) |
| F | Fidelity | yes (min under ⊗) |
| K | Kinetic character | — |
| G | Scope/granularity | — |
| Γ | Interaction grammar | — |
| Φ | Criticality | — |
| H | Chirality/temporal depth | — |
| S | Stoichiometry | — |
| Ω | Topological protection | — |
Union primitives (D, T, R, K, G, Γ, Φ, H, S, Ω) take max under tensor. Bottleneck primitives (P, F) take min — the weaker partner always wins. This is the structural enforcement mechanism behind the Frobenius non-synthesizability theorem.
| Tier | Condition | Letters |
|---|---|---|
| O_∞ | Φ_c + P_±^sym (Frobenius) | ו, מ, ש |
| O_2 | Φ_c + Ω ≠ Ω_0 + D ≠ D_∞ | א, ה, ע, ק, ת |
| O_1 | Φ_c + Ω = Ω_0 | ל |
| O_0 | Sub/super-critical | Remaining 13 |
Ker(I) = {(x,y) ∣ I(x) = I(y)} is a congruence on (𝒜, ⊗, ∨, ∧, med).
Proof: 0 failures in exhaustive sweep over all Ker(I) pairs × all operations × all contexts.
Consequence: λ_ℵ / Ker(I) is a well-defined 18-class quotient algebra.
The three Frobenius fixed points are pairwise I-distinguishable:
| Pair | Distance |
|---|---|
| d_I(ו, מ) | 14.92 |
| d_I(ו, ש) | 16.68 |
| d_I(מ, ש) | 4.84 |
No terminal object exists. Infinity is a relational structure, not a point.
For 18/22 letters z: d_I(med(z, מ, ש), מ) < d_I(z ⊗ מ, מ).
Mediation never loses globally. The 2-cell operation dominates the 1-cell.
22 boundary generators collapse to 18 behavioral classes. The 4 excess dimensions are structurally necessary — removing any canonical letter breaks the interaction structure.
α^(n)[med(ו, b, ש)] and α^(n)[med(ו, b', ש)] are α^(k)-equivalent for k ≤ n+2 and α^(k)-inequivalent for k ≥ n+3, where I(b) = I(b') but b ≠ b' syntactically.
Case 2 confirmed: λ_ℵ is not a quotient of any standard type theory.
d_I(x,y) = ∥v_x − v_y∥₂ exactly, where v_x ∈ ℝ²⁶⁴ is the weighted profile vector. The Gram matrix has rank 17. The interaction Hilbert space ℋ_I ≅ ℝ¹⁷ is a genuine inner product space.
Let G⁺ = {ג, ה, מ, [ב]} and G⁻ = {ס, ע, ש, [ד]}. Then for every h ∈ ℒ and every primitive k:
∑_{g ∈ G⁺} (g ⊗ h)k = ∑{g ∈ G⁻} (g ⊗ h)_k
Holds under ⊗, ∨, and ∧. All 264 primitive-by-primitive checks pass exactly. This is an exact algebraic theorem, not a metric property.
ק (Qoph, tier O_2) satisfies every O_∞ condition except P = P_±^sym. It is:
- The nearest non-Frobenius letter to מ: d_I(ק, מ) = 13.39 < d_I(ו, מ) = 14.92
- Interaction-row-equivalent to מ for 19/22 letters (differs only on {ו, מ, ש})
- A mediation gateway: med(ק, f, f') ∈ O_∞ for any f, f' ∈ Fix_∞
The grammar's central theorem states: Φ_c systems self-model — self-application reveals structure invisible at the definitional level. The grammar satisfies Φ_c. The interaction functor is the grammar's self-application. The Octad Balance, ק's position, and the rank-17 anomaly are exactly the class of discovery this theorem predicts.
Note
The grammar was correct about itself.
The operating system kernel is a single λ_ℵ term:
kernel = α[med(ו, מ ⊗ ש, □_Ω(א ⊗ (ש ⊗ מ)))]
| Component | λ_ℵ Operation |
|---|---|
| Process scheduling | Mediation |
| Memory allocation | Join (∨) |
| Inter-process communication | Tensor (⊗, P-bottlenecked) |
| Filesystem | The type lattice |
| Security | α-gating (C1–C4 coherence conditions) |
| Shell | λ_ℵ REPL (aleph_eval.py) |
| Boot | Tzimtzum: O_∞ → 22-letter alphabet → full environment |
Fundamental guarantee: ℵ-OS ⊗ -OS = ℵ-OS
The operating system is a Frobenius fixed point — idempotent under self-composition.
Each file represents a stage of discovery. Run them in order; each builds on the last.
Defines I(x) and d_I. Discovers the 4 equivalence collapses (22→18). Proves the interaction rows of ו, מ, ש are pairwise distinct despite all being O_∞.
Exhaustive substitutivity sweep: 0 failures. Ker(I) is a congruence. Mediation wins 18/22 over tensor at O_∞ proximity. Holographic interpretation established.
Constructs α[med(ו, ב, ש)] and α[med(ו, ח, ש)] with full history trees. Tests α^(n)-equivalence at depths 0–5. Case 2 confirmed at depth 4. Break-point law: α^(n) diverges at depth n+3.
Proves d_I is Euclidean. Constructs the Gram matrix. Finds rank 17 (not 18): one extra null dimension beyond Ker(I). Discovers the ק anomaly (φ_∞(ק) > φ_∞(ו)).
5️⃣ aleph_hidden_relation.py — What is the extra null direction?
Extracts the null eigenvector orthogonal to Ker(I). Identifies the Octad Balance: 4+4 perfect signed balance among 8 Hebrew letters, tier-symmetric. Proves it holds pointwise for all primitives.
A: Involution search — τ is not a permutation; concentrates at ל; Vav cast fails.
B: ק anatomy — one primitive from O_∞; mediation gateway; 19/22 row match with מ.
C: Axiom derivation — Octad Balance holds under ⊗, ∨, ∧; 792 checks; exact.
| Document | Purpose | Read if you want to... |
|---|---|---|
docs/ALEPH_SPEC.md |
Formal specification | Understand the calculus axiomatically (typing rules, reductions, C1–C4, §10 ℵ-OS) |
docs/LAMBDA_ALEPH.md |
Type theory reference | See the categorical model, collapse attack analysis, conditional univalence |
docs/ALEPH_DISCOVERY.md |
Narrative record | Follow the investigation from inception to completion |
docs/TECHNICAL_CONTRIBUTIONS.md |
Academic paper | Present the results to a mathematical audience |
docs/HEBREW_TYPE_LANGUAGE.md |
Alphabet encoding | See how each letter was assigned its 12-primitive tuple |
docs/PRIMITIVE_THEOREMS.md |
Formal theorem registry | Reference §23 (Frobenius non-synthesizability) and all prior theorems |
docs/SYNTHONICON_ONTICS.md |
Ontological grounding | Understand the broader SynthOmnicon framework |
docs/SYNTHONICON_DIAPHORICS.md |
Empirical predictions | See P-135/P-136 (Hebrew structural depth) |
docs/EGYPTIAN_MEDU.md |
Comparative alphabet | Medu Neter (hieroglyphics) as a second alphabet system |
- Normalization — Does λ_ℵ have a normal form theorem? Is reduction confluent?
- Full abstraction — If I(t₁) = I(t₂) in all contexts, does t₁ ≡ t₂ definitionally?
- *The -involution — τ concentrates at ל (O_1). Is there a tier-indexed involution giving L_{τ(x)} = L_x^†?
- Axiom proof of T7 — Explain why these 8 specific letters balance. What property of their primitive assignments forces the Octad Balance?
- ק's role — Is Qoph a designated O_∞ mediator in the process algebra? What operations require a threshold witness?
- Distributed ℵ-OS — Does α-gating survive network composition? Prove the idempotency guarantee holds across instances.
- Export — State the λ_ℵ axiom system purely mathematically, independent of the Hebrew encoding. Characterize the class of algebras satisfying T1–T8.
hott_bridge.py constructs the univalence bridge between the Hebrew lattice and Homotopy Type Theory.
Every letter in λ_ℵ has P ≤ P_sym. HoTT's identity type requires P_±^sym globally. The bridge is a single primitive lift:
d_HoTT = √w_P = √1.8 ≈ 1.3416
This is a near-grounded gap (just above √1, below √2) — the smallest possible structural separation.
| Operation | Method | Effect |
|---|---|---|
| Gap report | gap_report() |
Returns the divergent primitive and distance |
| System promotion | promote_to_hott() |
Clones alphabet with P → P_±^sym system-wide |
| Vav-cast | univalence_cast(a, b) |
Verifies d(a,b) < τ and lifts to HoTT identity |
The threshold τ is 4.0 for pairs with Ω ≥ Ω_{Z₂} (topologically protected), and 1.5 otherwise.
Note
Why Vav? ו (Vav, O_∞) is the unique letter whose interaction row is closest to the HoTT identity functor: P_±^sym, Φ_c, Ω_Z, T_⊙. The cast is named after it.
λ_ℵ is not a standard type theory, monoidal category, von Neumann algebra, or quotient of any existing framework. Proposed classification:
Aleph Coherence Geometry (ACG): a geometry in which objects are defined by their interaction profiles, equivalence is induced by behavioral indistinguishability, coherence paths (mediations) are the geodesics, and identity is a derived quotient of interaction structure.
Released under the MIT License.
The grammar was built on Φ_c. It found Φ_c in itself. The theorem proved itself.

