--- title: Residual Widom Factors in Extremal Polynomials url: https://www.emergentmind.com/topics/residual-widom-factors type: topic --- # Residual Widom Factors in Extremal Polynomials Residual Widom factors designate a family of normalization-dependent quantities in the theory of extremal polynomials. The common theme is removal of the dominant potential-theoretic scale—usually $\operatorname{Cap}(K)^n$, and in exterior-point problems $e^{-n g_K(x_0)}$—so that the remaining factor measures geometric, weighted, or asymptotic deviation from a baseline. The terminology is not standardized. In some papers it means an explicit nonnegative gap above a Szegő lower bound; in others it denotes Widom factors for residual polynomials; in still others it appears only as an interpretive description of geometric multipliers or of the remainder after automorphic normalization. Several recent works state explicitly that the exact phrase is not part of their formal definitions [2005.09114][2508.15131][2504.17727][2108.01798][1104.1915][2112.06450]. ## 1. Classical normalization and competing meanings The classical Widom normalization starts from a capacity-scaled extremal norm. For a compact set $E \subset \mathbb{R}$ and its monic Chebyshev polynomial $T_n$, the set-based Widom factor is \[ W_n(E):=\frac{\|T_n\|_E}{\operatorname{Cap}(E)^n}. \] For a compactly supported measure $\mu$ and $00}\frac{[W_{p,n}(\mu)]^p}{S(\mu)}=1. \] Thus the residual gap can be made arbitrarily small within the class of polynomial perturbations of the equilibrium measure. A factor-$2$ residual is therefore a phenomenon of additional structure, not of the basic Szegő inequality itself [2005.09114]. Generalized Jacobi measures provide a precise intermediate regime. If \[ d\mu(x)=(1-x)^\alpha(1+x)^\beta\,d\mu_K(x), \qquad \alpha,\beta\in\mathbb{N}\cup\{0\},\ \alpha+\beta\ge 1, \] on a regular compact $K\subset[-1,1]$ with $\pm1\in K$, then \[ [W_{2,n}(\mu)]^2\ge 2S(\mu). \] For the special cases $d\mu(x)=(1-x^2)d\mu_K(x)$, $d\mu(x)=(1-x)d\mu_K(x)$, and $d\mu(x)=(1+x)d\mu_K(x)$, vanishing of the residual \[ [W_{2,n}(\mu)]^2-2S(\mu) \] is characterized by inverse-image representations of $K$, and in the saturation regime the $L^2$ extremal polynomial coincides with the corresponding weighted Chebyshev polynomial [2107.13245]. ## 3. Residual polynomials and exterior-point Widom factors A second formal meaning arises from residual polynomials normalized at an exterior point. For a non-polar compact $K\subset\mathbb{R}$ and $x_0\in\mathbb{R}\setminus K$, the $n$th residual polynomial $R_{n,K}^{(x_0)}$ is the unique polynomial of degree at most $n$ minimizing the sup norm on $K$ subject to \[ R_{n,K}^{(x_0)}(x_0)=1. \] Its natural exponential scale is determined by the Green function: \[ \lim_{n\to\infty}\|R_{n,K}^{(x_0)}\|_K^{1/n}=e^{-g_K(x_0)}, \qquad \|R_{n,K}^{(x_0)}\|_K\ge e^{-n g_K(x_0)}. \] The residual Widom factor is therefore defined by \[ W_{\infty,n}^{(x_0)}(K):=\frac{\|R_{n,K}^{(x_0)}\|_K}{e^{-n g_K(x_0)}}\ge 1. \] It is the direct exterior-point analogue of the capacity-normalized Chebyshev factor [2508.15131]. This definition supports a strong unboundedness theory. Using weakly equilibrium Cantor sets $K(\gamma)$, one can prescribe subexponential lower growth. Given any sequence $(c_n)$ with subexponential growth, there exists a non-polar weakly equilibrium Cantor set such that \[ W_{2,n}(\mu_{K(\gamma)})\ge c_n\qquad\text{for all }n. \] For the same set and every exterior point $x_0\in\mathbb{R}\setminus K(\gamma)$, the residual factors satisfy \[ W_{\infty,n}^{(x_0)}(K(\gamma))\ge c_n^{\tau_{x_0}} \] in an unbounded gap, and \[ W_{\infty,2^s}^{(x_0)}(K(\gamma))\ge c_{2^s}^{\tau_{x_0}} \] along a dyadic subsequence in a bounded gap, where \[ \tau_{x_0}:=\frac{1}{\tau_{\Omega_{x_0}}(x_0,\infty)}\in(0,1]. \] If $(c_n)$ is monotone increasing and unbounded, then the residual Widom factors are unbounded for every exterior point. The lower bounds are derived from period-$n$ approximants, harmonic-measure representations for differences of Green functions, and Harnack inequalities [2508.15131]. The residual-polynomial viewpoint also extends to weighted complex-analytic settings. For a polynomially convex compact set $K\subset\mathbb{C}$ with connected complement and bounded weight $w$, the weighted residual polynomial $R_n^{K,w}(\cdot,z_0)$ is defined by a sup-norm extremal problem anchored at $z_0\in\overline{\mathbb{C}}\setminus K$, and the associated weighted Chebyshev polynomial is \[ T_n^{K,w}(z,z_0):=\frac{R_n^{K,w}(z,z_0)}{R_n^{K,w}(z_0,z_0)}. \] On the circular arc \[ \Gamma_\alpha=\{e^{it}:-\alpha\le t\le \alpha\}, \] the weighted Widom factors \[ \mathcal{W}_n(\Gamma_\alpha,w):=\frac{\|wT_n^{\Gamma_\alpha,w}\|_{\Gamma_\alpha}}{\operatorname{Cap}(\Gamma_\alpha)^n} \] satisfy \[ \lim_{n\to\infty}\mathcal{W}_n(\Gamma_\alpha,w) = 2\cos^2(\alpha/4)\, \exp\!\left(\int_{\Gamma_\alpha}\log w(x)\,d\mu_{\Gamma_\alpha}(x)\right). \] Here the residual scale is the explicit factor $2\cos^2(\alpha/4)$ multiplied by the arc Szegő integral [2602.05428]. ## 4. Arc geometry, asymmetric boundary behavior, and mixed arc–curve sets For a $C^{2+}$ Jordan arc $\Gamma$, residual behavior can be encoded by a geometric multiplier rather than by a difference. If $\mu=f\,d\mu_\Gamma$ is in the Szegő class, then \[ \lim_{n\to\infty}[W_{2,n}(\mu)]^2 = v(\mu_\Gamma)\,S(f), \] where \[ 1\le v(\mu_\Gamma)\le 2, \] and \[ v(\mu_\Gamma)=2 \quad\Longleftrightarrow\quad g'_+(z)=g'_-(z)\ \text{for all }z\in\Gamma_0. \] A natural residual geometry factor is \[ R_{\mathrm{res}(\Gamma)} := \sqrt{\pi\,R_{\mu_\Gamma}(\infty)\,\operatorname{cap}(\Gamma)} = \sqrt{\frac{v(\mu_\Gamma)}{2}} \le 1. \] In the sup norm, this yields the improved upper bound \[ \limsup_{n\to\infty}W_{\infty,n}(\Gamma,p) \le 2\,R_{\mathrm{res}(\Gamma)}\,S(p), \] and if there exists an interior point with $g'_+(z)\neq g'_-(z)$, then \[ \limsup_{n\to\infty}W_{\infty,n}(\Gamma,p)<2S(p). \] Non-analyticity of the open arc $\Gamma_0$ also forces strict improvement in both the $L^2$ and sup-norm settings [2108.01798]. The geometric meaning is explicit in model examples. For $\Gamma=[-1,1]$, one has $g'_+=g'_-$ on the whole interior, so $v(\mu_\Gamma)=2$, $\lim_{n\to\infty}[W_{2,n}(\mu_\Gamma)]^2=2$, and for $p\equiv1$ the weighted Chebyshev factors satisfy \[ W_{\infty,n}(\Gamma,1)=2 \quad\text{for all }n. \] For a circular arc, by contrast, \[ \lim_{n\to\infty}W_{\infty,n}(\Gamma,1)<2, \] which exhibits a strictly smaller residual multiplier [2108.01798]. A related but distinct arc phenomenon appears for compact sets consisting of smooth Jordan curves and arcs. In that setting, a naive universal “multiply by $2$” rule fails. If $E$ has at least one Jordan curve component, then \[ \limsup_{n\to\infty} \frac{M_{n,p}}{\operatorname{cap}(E)^n\,p(p,\mathcal{T}_n)} \le \theta<2, \] where $\theta=p(\tau,0)$ and $\tau$ doubles the weight on the arc components. In symmetric configurations consisting of real intervals together with Jordan curves symmetric with respect to $\mathbb{R}$, the sharp asymptotic is instead \[ M_{n,p}\sim \operatorname{cap}(E)^n\,p(p^*,\mathcal{T}_n), \] with the arc-adjusted weight $p^*$ equal to $p$ on curve components and $2p$ on arc components. Relative to this arc-adjusted envelope, the residual factor tends to $1$ [1401.6357]. These results correct a common misconception. The factor $2$ is exact for intervals and some symmetric arc problems, but it is not a universal residual constant for all arc-containing sets. ## 5. Multivariate formulations in $\mathbb{C}^n$ The multidimensional theory of Widom factors extends the normalization principle from $\mathbb{C}$ to $\mathbb{C}^n$, but it does not formally define residual Widom factors. For a compact non-pluripolar $K\subset\mathbb{C}^n$, with Monge–Ampère measure \[ dv_K:=\left(\frac{1}{2\pi}\right)^n (dd^cV_K)^n, \] the $L^2$ Widom factor attached to the monic orthogonal polynomial $V_{\alpha(i)}$ is \[ W_{2,\alpha(i)}(K,w):=\|V_{\alpha(i)}\|_{L^2(w\,dv_K)}, \] while the sup-norm factor attached to the weighted Chebyshev polynomial $T(K)_{\alpha(i),\widehat w}$ is \[ W_{\infty,\alpha(i)}(K,\widehat w) := \frac{\|T(K)_{\alpha(i),\widehat w}\|_K}{[T_-(K)]^{|\alpha(i)|}}. \] The normalizing role played by capacity in one variable is assumed here by the multidimensional Chebyshev constant $T_-(K)$ [2504.17727]. On product sets \[ K=K_1\times\cdots\times K_n, \qquad K_j\subset\mathbb{C}\ \text{non-polar}, \] the theory tensorizes. The Monge–Ampère measure factors as \[ dv_K=d\mu_{K_1}\otimes\cdots\otimes d\mu_{K_n}, \] and for product weights \[ w(x_1,\dots,x_n)=\prod_{j=1}^n w_j(x_j), \qquad S(K,w)=\prod_{j=1}^n S(K_j,w_j). \] This yields the lower bounds \[ [W_{2,\alpha(i)}(K,w)]^2\ge S(K,w), \qquad W_{\infty,\alpha(i)}(K,\widehat w)\ge S(K,\widehat w), \] which are direct multivariate analogues of one-dimensional Szegő inequalities. If each $K_j\subset\mathbb{R}$ and $w\equiv1$, the lower bounds improve to \[ [W_{2,\alpha(i)}(K,1)]^2\ge 2, \qquad W_{\infty,\alpha(i)}(K,1)\ge 2 \quad\text{for all }i\ge 1. \] Equality is characterized by inverse-image conditions of the form $K_j=R_{a_j}([-1,1])$ when the corresponding component degree $a_j$ is nonzero [2504.17727]. In this setting the paper explicitly states that “residual Widom factors” are not part of the formal theory. A natural interpretation is therefore residual-by-subtraction: \[ \mathrm{residual}_{2,\alpha(i)}(K,w):=[W_{2,\alpha(i)}(K,w)]^2-S(K,w), \] \[ \mathrm{residual}_{\infty,\alpha(i)}(K,\widehat w):=W_{\infty,\alpha(i)}(K,\widehat w)-S(K,\widehat w). \] These quantities are not defined as canonical invariants, but the inequalities above make them meaningful diagnostics of deviation from “Szegő-level” behavior. On real product sets with $w\equiv1$, both residuals are at least $1$ [2504.17727]. The same framework connects residual growth to Mahler measure. For any polynomial $P$, \[ \|P\|_{L^2(w\,dv_K)}^2 \ge S(K,w)\, \exp\!\left(\int_K \log|P|\,dv_K\right), \] so the excess above the Szegő baseline is constrained by the Mahler measure of $P$ relative to $K$ [2504.17727]. ## 6. Automorphic and Parreau–Widom residual structure In the finite-gap and Parreau–Widom literature, residual Widom behavior is often understood as the oscillatory remainder after removing both the capacity scale and the automorphic envelope. For Chebyshev polynomials on a compact set $E\subset\mathbb{C}$, \[ W_n(E):=\frac{\tau_n(E)}{\operatorname{Cap}(E)^n}, \] and the relevant oscillatory factor is supplied by Widom minimizers $F_{\chi}$ in automorphic Hardy spaces. Writing $F_n:=F_{\chi_E^n}$, the review of Christiansen, Simon, and Zinchenko describes the residual factor \[ R_n(E):=\frac{W_n(E)}{c(E)\,\|F_n\|_\infty}, \] with \[ c(E)=2 \quad\text{for real Parreau\text{–}Widom sets with DCT}, \qquad c(E)=1 \quad\text{for region-only weighted settings}. \] Then \[ R_n(E)\to 1. \] The sequence $\|F_n\|_\infty$ is periodic when the character is torsion and almost periodic otherwise, so the residual factor isolates the non-oscillatory part of the asymptotics [2112.06450]. A closely related decomposition appears in Christiansen’s treatment of Szegő’s theorem on Parreau–Widom sets. There the phrase “residual Widom factors” is not used, but the canonical factorization of the $M$-function separates geometric Blaschke factors, spectral Blaschke factors, and an outer factor determined by boundary modulus. The gap-critical values \[ W_j:=e^{-g_E(c_j)} \] may be viewed as Widom factors attached to the gaps, since the Parreau–Widom condition \[ \sum_j g(c_j)<\infty \] implies convergence of the associated product. After extracting these Blaschke contributions, the remaining residual component is purely outer and is governed by the Szegő integral \[ \int_E \log f\,d\mu_E. \] This residual outer term enters the step-by-step sum rules for \[ \frac{a_1\cdots a_n}{\operatorname{Cap}(E)^n}, \] and the absence of a singular inner part is one of the structural consequences of the Parreau–Widom condition [1104.1915]. This automorphic viewpoint suggests a general principle. Residual Widom factors are most stable when one distinguishes three layers: the leading potential-theoretic scale, the oscillatory automorphic or geometric envelope, and the final residual term. Different branches of the theory place the word “residual” at different layers, which explains the terminological variation across current work. Source: https://www.emergentmind.com/topics/residual-widom-factors