OctoInformBeta

OpenAI preprint, September 2026 · §6. Infinite paths and weak-Cauchy block sequences

§6 (setting): actual infinite paths and admissible blocks: Raw coding data index the paths, while real block sequences obey pure or mixed succession

Infinite paths retain their original pieces.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

Paths and block sequences have coordinates in the first uncountable ordinal Γ=ω1\Gamma=\omega_1. Fix a bijection π:Γ→ℕ×Γ\pi:\Gamma\longrightarrow\mathbb N\times\Gamma.✦ The color c(a)c(a) of a∈Γa\in\Gamma is the first component of π(a)\pi(a), and Ck={a∈Γ:c(a)=k}C_k=\{a\in\Gamma:c(a)=k\}.✦ Each color class has cardinality ℵ1\aleph_1 and is unbounded in ω1\omega_1.

For 𝕂=ℚ\mathbb K=\mathbb Q or ℝ\mathbb R, the space c00(Γ,𝕂)c_{00}(\Gamma,\mathbb K) consists of the 𝕂\mathbb K-valued functions on Γ\Gamma with finite support. Write 𝒜=c00(Γ,ℚ)\mathcal A=c_{00}(\Gamma,\mathbb Q) for the rational arrays.✦ The operation families are the pure families 𝗉k\mathsf p_k, k≥0k\geq0, and the mixed family 𝗆\mathsf m.✦ For finitely supported arrays u,vu,v, succession for a family ff is the relation u≺fvu\prec_f v defined by (1). u≺fv⇔{a<bfor all a∈supp⁡u,b∈supp⁡v,c(a)<c(b)for all such a,b when f=𝗆.u\prec_f v \quad\Longleftrightarrow\quad \begin{cases} a<b &\text{for all }a\in\operatorname{supp}u, \ b\in\operatorname{supp}v,\\ c(a)<c(b) &\text{for all such }a,b\text{ when }f=\mathsf m. \end{cases}(1)✦

Parameters and Type I Expressions

For a natural parameter ss, the depth sequence dnd_n and the parameter state (ts,n,bs,n)∈ℕ2(t_{s,n},b_{s,n})\in\mathbb N^2 are defined by (2); ts,nt_{s,n} is the exponent scale. dn=n+11,(ts,0,bs,0)=(d02+2d0+2,0),bs,n+1=bs,n+2⋅2s(s+1)ts,n,ts,n+1=ts,n+(bs,n+1+1)(dn+1+1)+1+dn+1+dn+12.\begin{aligned} d_n&=n+11,& (t_{s,0},b_{s,0})&=(d_0^2+2d_0+2,0),\\ b_{s,n+1}&=b_{s,n}+2\cdot2^{s(s+1)t_{s,n}},\\ t_{s,n+1}&=t_{s,n}+(b_{s,n+1}+1)(d_{n+1}+1) +1+d_{n+1}+d_{n+1}^2. \end{aligned}(2) All sequence indices begin at zero. The underlying block lengths and normalization factors, and their family versions, are given by (3). Here s𝗉k=k+2s_{\mathsf p_k}=k+2 and s𝗆=2s_{\mathsf m}=2; weight j≥1j\geq1 uses underlying parameter index j−1j-1.✦ Ls,n=2s(s+1)ts,n,ms,n=2(s−1)(s+1)ts,n+dn,Ljf=Lsf,max⁡{j−1,0},mjf=msf,max⁡{j−1,0}.\begin{aligned} L_{s,n}&=2^{s(s+1)t_{s,n}},& m_{s,n}&=2^{(s-1)(s+1)t_{s,n}+d_n},\\ L_j^f&=L_{s_f,\max\{j-1,0\}},& m_j^f&=m_{s_f,\max\{j-1,0\}}. \end{aligned}(3) The subtraction in the exponent is natural-number subtraction.

A Type I expression e=(j;z0,…,zℓ−1)e=(j;z_0,\ldots,z_{\ell-1}) for ff consists of a weight j≥1j\geq1 and nonzero children zh∈𝒜z_h\in\mathcal A, where 1≤ℓ≤Ljf1\leq\ell\leq L_j^f and zh≺fziz_h\prec_f z_i whenever h<ih<i.✦ Its value V(e)V(e) is the normalized rational sum in (4). V(e)=1mjf∑h=0ℓ−1zh.V(e)=\frac{1}{m_j^f}\sum_{h=0}^{\ell-1}z_h.(4)✦

The Norming Families

For a base set B⊆𝒜B\subseteq\mathcal A and an operation family ff, the norming construction stages Tn(B,f)⊆𝒜T_n(B,f)\subseteq\mathcal A start with T0(B,f)=BT_0(B,f)=B. At each subsequent stage, retain the preceding stage and adjoin the values of all Type I expressions whose children lie in that stage and all Type II expressions for which every child of every original uncropped piece on every original path lies in that stage.

The signed monochromatic coordinate base BkB_k, the pure norming set KkK_k, and the norming set KfK_f for a family ff are defined by (5); δa\delta_a denotes the coordinate array at aa. Bk={0}∪{δa,−δa:a∈Ck},Kk=⋃n≥0Tn(Bk,𝗉k),B𝗉k=Bk,B𝗆=⋃k≥0Kk,Kf=⋃n≥0Tn(Bf,f).\begin{aligned} B_k&=\{0\}\cup\{\delta_a,-\delta_a:a\in C_k\},& K_k&=\bigcup_{n\geq0}T_n(B_k,\mathsf p_k),\\ B_{\mathsf p_k}&=B_k,& B_{\mathsf m}&=\bigcup_{k\geq0}K_k,\\ K_f&=\bigcup_{n\geq0}T_n(B_f,f). \end{aligned}(5) The base norming set Bf⊆𝒜B_f\subseteq\mathcal A of a family ff consists of the signed monochromatic coordinates, including zero, when f=𝗉kf=\mathsf p_k, and is the union of the pure norming sets over all colors when f=𝗆f=\mathsf m.

Finite Closures and Raw Codes

A predecessor closure family GG at β∈Γ\beta\in\Gamma is a family (Gn(a))a<β,n≥0(G_n(a))_{a<\beta,\,n\geq0} of finite subsets of Γ\Gamma.✦ A predecessor closure family G=(Gn(a))a<β,n≥0G=(G_n(a))_{a<\beta,\,n\geq0} at β∈Γ\beta\in\Gamma is coherent when a∈Gn(a)⊆[0,a]a\in G_n(a)\subseteq[0,a], the sets at each aa increase to [0,a][0,a], and Gn(b)=Gn(a)∩[0,b]G_n(b)=G_n(a)\cap[0,b] whenever b∈Gn(a)b\in G_n(a).✦ A coherent endpoint extension (Hn)(H_n) of a predecessor closure family GG at β\beta is a sequence of finite subsets of [0,β][0,\beta] containing β\beta, increasing to [0,β][0,\beta], and satisfying Gn(a)=Hn∩[0,a]G_n(a)=H_n\cap[0,a] whenever a<βa<\beta belongs to HnH_n.✦

Write Nβ(G)N_\beta(G) for the endpoint construction, which assigns a sequence of finite subsets of Γ\Gamma to a predecessor closure family at β\beta. The coherent finite closure system RR is defined by ordinal recursion in (6). The sets Fp(β)F_p(\beta) use positive closure indices, and ρ(α,β)\rho(\alpha,\beta) is the closure entry rank. R•(β)=Nβ((R•(a))a<β),Fp(β)=Rmax⁡{p−1,0}(β),ρ(α,β)={min⁡{p≥1:α∈Fp(β)},α≤β,0,α>β.\begin{aligned} R_\bullet(\beta)&=N_\beta\bigl((R_\bullet(a))_{a<\beta}\bigr),\\ F_p(\beta)&=R_{\max\{p-1,0\}}(\beta),\\ \rho(\alpha,\beta)&= \begin{cases} \min\{p\geq1:\alpha\in F_p(\beta)\},&\alpha\leq\beta,\\ 0,&\alpha>\beta. \end{cases} \end{aligned}(6)

A raw infinite path r=(ui,ji,pi)i≥0r=(u_i,j_i,p_i)_{i\geq0} consists of nonzero rational arrays ui∈𝒜u_i\in\mathcal A, declared weights ji∈ℕj_i\in\mathbb N, and metadata pi∈ℕp_i\in\mathbb N.✦ Its prefix support Sn(r)S_n(r), endpoint βn(r)\beta_n(r), and prefix closure Dn(r)D_n(r) are the finite-coordinate constructions in (7). Sn(r)=⋃i=0nsupp⁡ui,βn(r)=max⁡Sn(r),Dn(r)=Fpn(βn(r)).\begin{aligned} S_n(r)&=\bigcup_{i=0}^n\operatorname{supp}u_i,\\ \beta_n(r)&=\max S_n(r),& D_n(r)&=F_{p_n}(\beta_n(r)). \end{aligned}(7)

A finite path code cc is a finite list of triples (j,p,A)(j,p,A), with j,p∈ℕj,p\in\mathbb N and AA a finite subset of ℕ×ℚ\mathbb N\times\mathbb Q.✦ For a finite coordinate set D⊆ΓD\subseteq\Gamma, the coordinate order position pos⁡D(a)\operatorname{pos}_D(a) and the positional rational array code AD(u)A_D(u) are defined by (8), retaining zero coefficients. pos⁡D(a)=|{b∈D:b<a}|,AD(u)={(pos⁡D(a),u(a)):a∈D}.\operatorname{pos}_D(a)=|\{b\in D:b<a\}|, \qquad A_D(u)=\{(\operatorname{pos}_D(a),u(a)):a\in D\}.(8) The prefix code cn(r)c_n(r) records the original pieces relative to the common prefix closure, as in (9). cn(r)=[(ji,pi,ADn(r)(ui))]i=0n.c_n(r)=\bigl[(j_i,p_i,A_{D_n(r)}(u_i))\bigr]_{i=0}^n.(9)✦

For a finite code cc, let M(c)M(c) be its maximum metadata, with M([])=0M([])=0.✦ The fixed injective coding map σ\sigma pairs the canonical natural-number code of cc with M(c)+1M(c)+1.✦ Its value exceeds every metadata entry of cc.

An admissible raw infinite path rr is a raw path with successive supports in ordinal order and positive, strictly increasing metadata, subject to (10). 1≤ji≤pi(i≥0),ρ(a,b)≤pn(a,b∈Sn(r),a≤b),jn+1=σ(cn(r))(n≥0).\begin{aligned} 1&\leq j_i\leq p_i &&(i\geq0),\\ \rho(a,b)&\leq p_n &&(a,b\in S_n(r),\ a\leq b),\\ j_{n+1}&=\sigma(c_n(r)) &&(n\geq0). \end{aligned}(10)✦

The Definition

Definition 1. Fix a pure or mixed family ff. An infinite path P∈𝒫fP\in\mathcal P_f is a pair (r,(ei)i≥0)(r,(e_i)_{i\geq0}), where r=(ui,ji,pi)i≥0r=(u_i,j_i,p_i)_{i\geq0} is a raw infinite path and ei=(wi;zi,0,…,zi,ℓi−1)e_i=(w_i;z_{i,0},\ldots,z_{i,\ell_i-1}) are Type I expressions for ff satisfying (11). r is admissible,V(ei)=ui,wi=ji(i≥0),zi,h∈Kf(i≥0,0≤h<ℓi),ui≺fuj(i<j).\begin{aligned} r&\text{ is admissible},\\ V(e_i)&=u_i,&w_i&=j_i &&(i\geq0),\\ z_{i,h}&\in K_f &&&&(i\geq0,\ 0\leq h<\ell_i),\\ u_i&\prec_f u_j &&&&(i<j). \end{aligned}(11)✦

The set ℐf\mathcal I_f of raw path indices for ff is the set of raw infinite paths that occur in infinite paths, as specified by (12). ℐf={r:there exists (ei)i≥0 with (r,(ei)i≥0)∈𝒫f}.\mathcal I_f=\{r:\text{there exists }(e_i)_{i\geq0} \text{ with }(r,(e_i)_{i\geq0})\in\mathcal P_f\}.(12)✦

An admissible block sequence (xn)n≥0(x_n)_{n\geq0} for ff is a sequence in c00(Γ,ℝ)c_{00}(\Gamma,\mathbb R) satisfying (13). xn≠0(n≥0),xi≺fxj(i<j),supp⁡xn⊆Ck(f=𝗉k,n≥0).\begin{aligned} x_n&\ne0 &&(n\geq0),\\ x_i&\prec_f x_j &&(i<j),\\ \operatorname{supp}x_n&\subseteq C_k &&(f=\mathsf p_k,\ n\geq0). \end{aligned}(13)✦

What Is Checked

The definitions concern the pure and mixed norming families on ω1\omega_1, with rational Type I pieces and finitely supported real block vectors. Every raw infinite path has nonempty prefix support at each index.✦