embed_add
plain-language theorem explainer
The embedding of LogicNat into positive reals preserves addition as multiplication for any generator γ. Researchers deriving arithmetic from the law of logic cite this when showing that iteration counts act as exponents in the orbit. The proof is by induction on the second argument, reducing the successor case with the recursive definitions of addition and embedding followed by ring simplification.
Claim. Let γ be a generator and let a, b be elements of the logic-induced naturals. Then the embedding of a + b equals the product of the embeddings of a and b.
background
LogicNat is the inductive type with constructors identity (zero-cost element) and step (one iteration), forming the smallest subset of positive reals closed under multiplication by γ.value and containing 1. The embed function is defined recursively: identity maps to 1 and step n maps to γ.value times the embedding of n. Addition on LogicNat satisfies add_zero (n + zero = n) and add_succ (n + succ m = succ (n + m)). A generator is any positive real other than 1, extracted from the non-triviality clause of the law of logic.
proof idea
The proof is by induction on b. The identity case rewrites via add_zero, embed_zero, and mul_one. The step case rewrites via add_succ and embed_succ on both sides, applies the inductive hypothesis, and closes with ring.
why it matters
This homomorphism shows that LogicNat is precisely the orbit parameterized by iteration count, completing the structural identification of naturals with multiplicative powers. It feeds the arithmetic layer of the foundation module and supports later derivations of the phi-ladder and mass formula. In the T0-T8 chain it supplies the discrete counting structure required before forcing D = 3 and the eight-tick octave.
Switch to Lean above to see the machine-checked source, dependencies, and usage graph.
papers checked against this theorem (showing 30 of 61)
-
Perturbation around Koebe function closes Li-Keiper equations
"λ(n) = ∑ (-1)^{k-n+1} binom(n,k) λ(k) + Δ (Eq. 12); λ_n = (-1)^n ∑ (-1)^k binom(2n,n-k) λ_k (Eq. 18); discrete derivative Δ^n f(m) = ∑ (-1)^k binom(n,k) f(m+(n-k)) (Appendix 3)"
-
Logic layers let neural nets learn rules and arithmetic
"x∧y = xy, x∨y = x + y − xy, ¬x = 1−x (eqs. 1-3); AND-rule and OR-rule via exp(A(log(|ˆx|+ϵ))) and Kronecker form (eqs. 12-13)"
-
Jaccard closest pair needs quadratic time for close thresholds
"Squaring: a'_ij = a_i · a_j; J after i squarings = |a∩b|^{2i} / (|a|^{2i}+|b|^{2i}-|a∩b|^{2i}) (Eq. 2, Lemma 7)"
-
Graph on top-k logits lifts fused LLM reasoning scores
"Dynamic Graph Construction... adjacency matrix Cb(i,j)=∑t zt(i)·zt(j)"
-
Modified deck-of-cards process builds interval type-2 fuzzy sets
"Theorem 2 (Marchant, 2004) ... multiplicative property ... μlr(x)/μlr(y)=ρlr(x|y)"
-
One speaker creates multiple sound zones with multi-frequency ultrasound
"pa(r, ωa) = Σ wn Ha,n(r, ωa) ... linear combination ... equivalently interpreted as N audio channels being virtually created in the air"
-
Tensor ranks control optimal rates for multivariate intensity estimation
"we show that many fundamental classes of multivariate functions, including additive and mean-field models, admit finite-rank tensor representations"
-
Right translation aligns arbitrary canonicals for symmetry discovery
"Proposition 3.1 ... the empirical distribution ˆµ[x] corresponding to the normalized samples ψ([x])ˆΓ⁻¹[x] approximates the target distribution µ[x] ... right-multiplying by the inverse ˆΓ⁻¹[x] corrects this offset and centers the Fréchet mean at the identity"
-
Exact Shapley values for product kernels in quadratic time
"Definition 1. ... νxxx(S) = α⊤ kS(XS, xxxS). ... removing a feature replaces its kernel factor with the multiplicative identity."
-
LMT pulse sequences prepare tunable quantum backflow states
"The free evolvement of a wavepacket between laser pulses can be described in a moving frame by a Galileo frame-transformation operator Ĝ"
-
Representation Gap is governed by task intrinsic dimension
"equivariant model virtually augments the dataset … Ωf = G(D)"
-
Reed-Muller RM(s-1,m) masks correct s stuck-at defects
"RM(r, m) = RM(r, m-1) | RM(r-1, m-1) (Plotkin construction)"
-
Message passing makes GNN-LRP subgraph scores linear in depth
"sGNN-LRP rule (4) only differs from the GNN-LRP rule (2) by introducing an additional summation that pools the propagated relevances over all nodes belonging to the subgraph S."
-
Shared bias removal aligns opposing ultrasonic backscatter profiles
"The method exploits the reciprocal constraint that the dominant through-thickness propagation bias should contain a shared component between opposing inspection directions"
-
Calibrated distillation lifts low-complexity speech enhancement
"recursive fusion to form the fused feature set that enables inter-set knowledge interaction"
-
Ergodic theorem links wavelets to Gaussian processes for AP-frames
"lim N→∞ 1/(2N+1) ∑ |⟨X,ψ_{j,k}⟩|^2 = E|⟨X,ψ_{j,0}⟩|^2 a.s."
-
Decentralized gossip approximates Wasserstein barycenters
"v ← (∏ K⊤ui)1/N ; log v = (1/N) ∑ si"
-
NMF incorporates feedback to separate direct and amplified effects
"The equilibrium operator M_model=(I−XΘ₁)⁻¹XΘ₂ admits the Neumann expansion … each term corresponds to additional rounds of latent propagation."
-
OMA retains Kubernetes crash evidence past the evidence horizon
"The causal edges OMA constructs are analogous to happened-before relationships [9]: if an OOMKillEvidence event e2 is observed for pod P within 90 seconds of an OOMKill event e1 for the same pod, then e1 → e2 in the happened-before sense."
-
Galois-field memory unifies hippocampus and causal reasoning
"path-integral Confidence Ratio CR2(n) = ∏ CR1(i) ... multiplicative decay"
-
Shapley values in product games collapse to one integral
"product game v(S) = ∏_{j∈S} u_j"
-
Input stacking extends vector fitting to non-square MIMO modal tests
"the sum of FRFs between inputs is equal to an effective FRF ... the summation does not affect the common poles since they are invariant under linear combinations"
-
ScheduleFree+ beats WSD schedules on long LLM pretraining
"Schedule-Free Learning provides a theoretical foundation for the use of model averaging and checkpoint merging during pretraining."
-
Unique Games admit tolerant testers with sublinear queries
"Lemma 3.11 (Trace ambiguity implies geometric ambiguity) ... uses Kac’s formula and geometric memorylessness."
-
Recycling upper-tail prices drives market price drift
"vt+T(h) = v_t(h) · r̄_t ... r̄_t = average of min winning ratios"
-
Trajectories cut video tokens by 10x while raising accuracy
"organizes tokens based on panoptic sub-object trajectories rather than fixed patches... Rooted in Spelke’s core cognitive principles and the Gestalt Principle of common fate"
-
High-dimensional equivalent extends RMT to nonlinear deep nets
"Theorem 8 … CK matrix Kℓ … ˜Kϕ,ℓ = α²ℓ,1 X⊤X + … recursion αℓ,1 = aϕℓ;1 · αℓ−1,1"
-
Algorithm simulates distributions with O(log^{2} N / ε^{2}) conditional samples
"repeated chain rule for DKL ... DKL(μ ∥ τ) = Σ ... marginals"
-
Area condition on f and g decides existence for p-Laplacian solutions
"integrating factor φ(u)=exp(−∫ g) and divergence-theorem identity after integration over Ω"
-
Robust LLM alignment reduces to relative reward regression with O(d/n) rates
"ℓREBEL(z;θ) = (1/η [log πθ(a1|x)/πt(a1|x) - ...] - [r(x,a1)-r(x,a2)])²"