Papers
Topics
Authors
Recent
Search
2000 character limit reached

Borel Complexity of Normal Vectors

Updated 7 January 2026
  • Borel complexity computations are defined by sets that are Pi^0_3-complete, expressed as countable intersections of F_sigma sets with continuous reductions from any such set.
  • The study employs recurrence sequences, Pisot bases, and symbolic dynamics to show that normality corresponds to uniform distribution mod 1 and is encoded in bi-infinite subshifts.
  • These findings extend classical base‑b normal numbers to higher-dimensional systems, providing deep insights into descriptive set theory and the structure of recurrence relations.

A set in descriptive set theory is Π30\Pi^0_3-complete if it can be expressed as a countable intersection of FσF_\sigma (i.e., countable unions of closed sets) subsets, and if every other Π30\Pi^0_3 set can be continuously reduced to it. The Borel complexity of sets of vectors normal for a fixed recurrence sequence provides deep insight into the structure of normality in extended numeration and dynamical regimes, connecting classical normal number sets with higher-dimensional recurrence behavior.

1. Normality in Recurrence Sequences and Companion Polynomials

For a recurrence sequence xn=ξ1α1n+ξ2α2nx_n = \xi_1 \alpha_1^n + \xi_2 \alpha_2^n determined by the companion polynomial P(X)P(X) (e.g., P(X)=(Xα1)(Xα2)P(X)=(X-\alpha_1)(X-\alpha_2)), the notion of a vector (ξ1,ξ2)(\xi_1, \xi_2) being "normal" means that the sequence (xn)n0(x_n)_{n \ge 0} is uniformly distributed modulo $1$: for every interval [a,b)[1/2,1/2)[a, b) \subset [-1/2,1/2),

limN1N#{0n<Nxn(mod1)[a,b)}=ba.\lim_{N \to \infty} \frac{1}{N} \#\left\{ 0 \leq n < N \mid x_n \pmod 1 \in [a, b) \right\} = b-a.

The set of normal vectors for PP is NP={(ξ1,ξ2)R2(xn) u.d. mod 1}N_P = \{ (\xi_1, \xi_2) \in \mathbb{R}^2 \mid (x_n) \text{ u.d. mod } 1 \} (Kaneko et al., 27 Oct 2025). Under mild algebraic hypotheses (notably: no root on the unit circle), this set is Borel and specifically Π30\Pi^0_3-complete.

2. Statement of Main Complexity Results

The principal theorem ((Kaneko et al., 27 Oct 2025), Theorem 1.1-1.2) states:

  • For every k0k \ge 0 and a fixed recurrence polynomial P(X)P(X) without roots on the unit circle, the set

NP,k={gΞk:(xn(g)) is u.d. mod 1}N_{P,k} = \{ \mathbf{g} \in \Xi_k : (x_n(\mathbf{g})) \text{ is u.d. mod } 1 \}

is Π30\Pi^0_3-complete in the Polish space Ξk\Xi_k parameterizing initial values.

  • Special case: If α|\alpha| is Pisot (α>1|\alpha| > 1 and all conjugates <1<1 in modulus), then for

Nα={ξR:(ξαn)n0 is u.d. mod 1}N_\alpha = \{ \xi \in \mathbb{R} : (\xi \alpha^n)_{n \ge 0} \text{ is u.d. mod } 1 \}

the set NαN_\alpha is Π30\Pi^0_3-complete, generalizing classical base-bb normal numbers (Ki–Linton (Airey et al., 2018)) to all Pisot bases (Kaneko et al., 27 Oct 2025).

3. Methodology: Symbolic Encoding and Subshifts

A central idea is encoding recurrence sequences' fractional parts by infinite words over a suitably chosen alphabet A\mathcal{A} (of size $2B+1$, with BB depending on PP's coefficients). Each sequence (xn)(x_n) can be recovered from

s=(sn)nZ,snA,s = (s_n)_{n \in \mathbb{Z}}, \quad s_n \in \mathcal{A},

via a convolution kernel explicitly defined in terms of PP (Kaneko et al., 27 Oct 2025), Propositions 2.1–2.2.

This symbolic correspondence realizes a bi-infinite subshift ΣPAZ\Sigma_P \subset \mathcal{A}^\mathbb{Z}, encoding all recurrence sequences. Normality (u.d. mod 1) for a vector g\mathbf{g} translates to the genericity of the associated s=Ψ(g)s = \Psi(\mathbf{g}) in ΣP\Sigma_P: every finite block ww must appear in ss with the proper "generic" frequency determined by Lebesgue measure.

The subshift ΣP\Sigma_P arising for PP is proved to have the right feeble specification property, a minimal form of mixing sufficient to guarantee that the set of generic points is Π30\Pi^0_3-complete (Airey et al., 2018, Kaneko et al., 27 Oct 2025).

4. Descriptive Complexity and Completeness Arguments

The Π30\Pi^0_3 membership of NPN_P is shown via cylinder-block frequency conditions:

  • For each block ww and tolerance ϵ>0\epsilon > 0, the set of ss whose empirical frequency approaches the limit within ϵ\epsilon is FσF_\sigma.
  • "Normality" requires these block-frequency conditions hold for all ww and ϵ\epsilon; thus NPN_P is a countable intersection of FσF_\sigma sets, i.e., Π30\Pi^0_3.

Π30\Pi^0_3-hardness follows from a Wadge reduction:

  • The standard Π30\Pi^0_3-complete set C3={βNN:lim infβ(n)}C_3 = \{ \beta \in \mathbb{N}^\mathbb{N} : \liminf \beta(n) \to \infty \} in the Baire space.
  • A continuous coding F:NNΞkF : \mathbb{N}^{\mathbb{N}} \to \Xi_k is constructed so that F(β)F(\beta) is normal (in the sense above) if and only if β(n)\beta(n) \to \infty.
  • The construction uses concatenations of "good" and "zero" blocks whose lengths depend on β(n)\beta(n); long "good" blocks drive the sequence towards genericity, while bounded β(n)\beta(n) prevent genericity (Kaneko et al., 27 Oct 2025), Lemmas 4.1–4.3.

5. Relation to Classical Normal Numbers and Broader Impact

This work subsumes the Ki–Linton (Airey et al., 2018) result that the set of base-bb normal numbers is Π30\Pi^0_3-complete and extends it to:

  • Numbers normal in any Pisot base,
  • Vectors normal for arbitrary recurrence relations.

For general numeration systems or continued fractions, symbolic codings yield subshifts of specification type, ensuring Π30\Pi^0_3-completeness of the normal set (Airey et al., 2018).

The methodology applies broadly: any system where normality corresponds to genericity for an invariant measure on a subshift with right feeble specification property will have the normal set Π30\Pi^0_3-complete.

6. Open Problems and Further Directions

Areas for further research include:

  • Extending results to non-Pisot bases (where digit expansions lack specification),
  • Joint normality for multiple recurrence sequences, involving higher complexity classes such as D2(Π30)D_2(\Pi^0_3),
  • Finer descriptive complexity above Π30\Pi^0_3 for systems lacking strong mixing or with more irregular limiting behavior.

7. Table: Main Π30\Pi^0_3 Normal Sets by Setting

System Object(s) Completeness Level Reference
Base-bb normal {ξ(ξbn) u.d. mod 1}\{\xi \mid (\xi b^n) \text{ u.d. mod } 1\} Π30\Pi^0_3-complete (Airey et al., 2018)
Pisot normal {ξ(ξαn) u.d. mod 1}\{\xi \mid (\xi \alpha^n) \text{ u.d. mod } 1\} (α|\alpha| Pisot) Π30\Pi^0_3-complete (Kaneko et al., 27 Oct 2025)
Recurrence normal {(ξ1,ξ2)xn=ξ1α1n+ξ2α2n u.d. mod 1}\{ (\xi_1, \xi_2) \mid x_n = \xi_1 \alpha_1^n + \xi_2 \alpha_2^n \text{ u.d. mod } 1 \} Π30\Pi^0_3-complete (Kaneko et al., 27 Oct 2025)

References

  • H. Kaneko, B. Mance, "Borel Complexity of the set of vectors normal for a fixed recurrence sequence" (Kaneko et al., 27 Oct 2025)
  • D. Airey, S. Jackson, D. Kwietniak, B. Mance, "Borel complexity of sets of normal numbers via generic points in subshifts with specification" (Airey et al., 2018)
  • H. Ki, T. Linton, "Normal numbers and subsets of N\mathbb{N} with given densities," Fund. Math. 144 (1994), 163–179.

In summary, for broad classes of recurrence and numeration systems—including all Pisot bases—the set of normal vectors or numbers is precisely Π30\Pi^0_3-complete in the Borel hierarchy, as established by Kaneko–Mance (Kaneko et al., 27 Oct 2025) and related works. This provides a robust uniform answer for normality in analytic and dynamical settings, marking a boundary in the complexity of natural number-theoretic properties.

Topic to Video (Beta)

No one has generated a video about this topic yet.

Whiteboard

No one has generated a whiteboard explanation for this topic yet.

Follow Topic

Get notified by email when new papers are published related to Borel Complexity Computations.