Truncated Seminorms: Theory & Applications

• Truncated seminorms are functionals that measure local smoothness by restricting difference interactions to neighborhoods scaled by the distance to the boundary.
• They enable interpolation between fractional Sobolev, Besov, and classical spaces, offering a flexible framework for analyzing nonlocal PDEs and variational problems on irregular domains.
• Their key properties—such as monotonicity, local-to-global equivalence, and controlled scaling—support critical results like fractional Korn-type inequalities and robust weighted estimates.

A truncated seminorm is a functional used to quantify the local or nonlocal smoothness of functions by restricting interactions—such as difference quotients or kernels—to neighborhoods whose size is typically controlled by proximity to the domain boundary. This localization, achieved through a truncation parameter, yields function spaces that intermediate between global (fractional) Sobolev, Besov, and related scales and their more classical counterparts. Truncated seminorms have become fundamental in contemporary analysis on general domains, including the study of nonlocal PDEs, variational problems with boundary effects, fractional Korn-type inequalities, and the regularization theory of jump processes. Their application and theory are especially prominent in uniform and John domains, and in the development of screened scales via truncated real interpolation.

1. Definitions and Foundational Types

Let $\Omega\subset\mathbb{R}^n$ be a bounded domain, with $\delta(x)=\mathrm{dist}(x, \partial\Omega)$. Common forms of truncated seminorms include:

• Truncated Gagliardo–Slobodeckij seminorm:

$$|u|_{W^{s,p}_\tau(\Omega)^n} := \left( \int_\Omega \int_{B(x, \tau \,\delta(x))} \frac{|u(x)-u(y)|^p}{|x-y|^{n+sp}}\,dy\,dx \right)^{1/p}$$

where $0$1<p<\infty$, and $\tau\in(0,1)$ controls the truncation radius.

• Truncated Mengesha (nonlocal symmetric gradient) seminorm:

$$|u|_{X^{s,p}_\tau(\Omega)^n} := \left( \int_\Omega \int_{B(x, \tau\,\delta(x))} \frac{|\,(u(y)-u(x))\cdot(y-x)|^p}{|y-x|^{n+sp+p}}\,dy\,dx \right)^{1/p}$$

• Truncated Triebel–Lizorkin/Sobolev-type seminorms (Rutkowski, 2018):

$$S^{K,\theta}_{p,q}(f;\Omega) = \left( \int_\Omega \left( \int_{|x-y|<\theta\,\delta(x)} |f(x)-f(y)|^q K(x,y)\,dy \right)^{p/q} dx \right)^{1/p}$$

for a general kernel $K(x,y)$, $1<q\le p<\infty$, and truncation parameter $\theta\in(0,1]$.

• Screened Besov and Sobolev seminorms (Stevenson et al., 2020) (on all of $\mathbb{R}^n$, truncation by maximum allowed difference $|h|<T$):

$$\|f\|_{\widetilde{B}^{s,p}_q} = \left(\int_{|h|<T} |h|^{-n-sp} \|f(\cdot+h)-f(\cdot)\|^q_{L^p} dh \right)^{1/q}$$

The special case $p=q$ yields the screened Sobolev seminorm.

These definitions encapsulate the restriction of nonlocal interactions to local neighborhoods, producing an analytic framework adapted to domains and boundary effects not well-handled by classical, globally-integrating seminorms.

2. Structural Properties and Comparisons

Truncated seminorms exhibit several structural and functional-analytic properties central to their utility:

• Scaling: For a cube $U$ of side-length $\ell$, under dilation $u_U(x)=u(\ell x)$, one has $$|u|_{W^{s,p}(U)} = \ell^{n/p-s} |u_U|_{W^{s,p}(\hat U)}$$, and analogously for truncated versions.
• Monotonicity: Truncating the radius of integration strictly decreases or coincides with the corresponding full seminorm, i.e., $$|u|_{W^{s,p}_\tau(\Omega)} \le |u|_{W^{s,p}(\Omega)}$$.
• Local-to-Global Equivalence: On uniform domains (domains in which every pair of points can be joined by a chain of balls with controlled overlap and interior localization), full and truncated seminorms are equivalent up to multiplicative constants depending on $n,p,s,\tau$ and the geometry of $\Omega$ (Acosta et al., 13 Jan 2026, Rutkowski, 2018):

$$|u|_{W^{s,p}(\Omega)} \simeq |u|_{W^{s,p}_\tau(\Omega)}$$

• Proper Interpolation Spaces: The truncated seminorm spaces $W^{s,p}_\tau(\Omega)$ sit strictly between $L^p$ and $W^{1,p}$ when $\Omega$ is not uniform, and naturally interpolate between local and global regularity (Acosta et al., 13 Jan 2026).
• Control by Gradients: For $1<p<\infty$, $|u|_{W^{s,p}_\tau(\Omega)}$ is always controlled by $\|\nabla u\|_{L^p(\Omega)}$.

On general (e.g., John) domains, truncated seminorms remain finite for a broader class of functions than their full analogues, crucial for analysis in rough domains or those with complicated boundaries.

3. Key Inequalities and Theoretical Outcomes

Truncated seminorms facilitate key analytic results, notably:

• Fractional Korn-type Inequalities: A novel approach employing truncated seminorms establishes a Korn-type inequality for fractional orders in John and uniform domains (Acosta et al., 13 Jan 2026). Specifically, for $u\in W^{s,p}_{\tau_2}(\Omega)^n$,

$$\inf_{r\in \mathrm{RM}} |u - r|_{W^{s,p}_{\tau_1}(\Omega)^n} \leq C\,|u|_{X^{s,p}_{\tau_2}(\Omega)^n}$$

where $\mathrm{RM}$ denotes rigid motions. This outcome generalizes previously domain-restricted results to non-smooth settings.

• Equivalence of Full and Truncated Seminorms in Uniform Domains: When $\Omega$ is uniform, both the fractional Korn inequality and Sobolev-type estimates hold with the global (full) seminorm replaced by its truncated version without analytic loss (Rutkowski, 2018).
• Weighted Inequalities: Introducing distance-to-boundary weights $d(x,y) = \min\{\delta(x), \delta(y)\}$, weighted truncated seminorms

$$|u|_{W^{s,p}_\tau(\Omega, d^{p\beta})}^p = \int_\Omega \int_{|x-y| < \tau \delta(x)} \frac{|u(x) - u(y)|^p}{|x - y|^{n+sp} d(x,y)^{p\beta}}\,dy\,dx$$

admit corresponding truncated Korn-type inequalities for suitable $\beta$ as determined by the Assouad dimension of the boundary (Acosta et al., 13 Jan 2026).

4. Methodological Principles: Truncation via Domain Geometry and Interpolation

The truncation mechanisms in seminorm construction reflect both domain geometry and analytical objectives:

• Boundary-Proportional Truncation: In canonical models, truncation is performed via $B(x, \tau\,\delta(x))$, localizing nonlocal interactions and mitigating the impact of singularities or domains lacking extension properties (Acosta et al., 13 Jan 2026, Rutkowski, 2018).
• Whitney Decomposition and Local-Global Patching: Uniform and John domains admit Whitney decompositions into cubes whose distances to the boundary are matched to their sizes. One may then apply local estimates on these cubes, inflating to smooth "cube" neighborhoods, then synthesize global control via tree-based discrete Poincaré inequalities and averaging arguments.
• Truncated Real Interpolation: The truncated $K$-method generates a scale of intermediate spaces via

$$\|x\|_{(X_0,X_1)_{s,q}^\sigma} = \left( \int_{0}^{\sigma} [t^{-s} K(t,x)]^q \frac{dt}{t} \right)^{1/q}$$

Interpolating $L^p$ and $\dot{W}^{1,p}$ under this truncated scheme yields the screened Besov and Sobolev spaces, characterized by modulation at scale $T$ and sum decompositions that reflect both high- and low-frequency content (Stevenson et al., 2020).

5. Case Studies: Successes, Failures, and Applications

Empirical and theoretical results establish contexts in which truncated seminorms are or are not comparable to their full versions:

Domain/Kernel Class	Truncated ≍ Full?	Typical Outcome
Uniform, fractional-power kernel	Yes	Truncation at $\theta\delta(x)$ causes no analytic loss (Rutkowski, 2018)
Non-uniform (thin strip), small-$\alpha$	No	Full and truncated norms can be vastly different
Integrable kernel (e.g., $K=1$)	No	Truncated seminorm can be arbitrarily small for large global effects
Dirichlet forms / Hunt processes	Yes (under suitable structure)	Analytic domains coincide for truncated and censored processes

Key applications include:

• Nonlocal PDEs and Peridynamics: Energy forms with finite interaction radius correspond to truncation, and truncated seminorms define function spaces on such models (Rutkowski, 2018).
• Fractional Korn Inequalities for Fractal Boundaries: On domains with Koch snowflake-type or fractal boundaries, truncated seminorms ensure coercivity and well-posedness where traditional methods fail (Acosta et al., 13 Jan 2026).
• Interpolation Theory and Trace Spaces: Truncated seminorms provide precise machinery for analyzing trace and embedding results in nonlocal and boundary-focused settings (Stevenson et al., 2020).

6. Topological and Sum-Decomposition Characterizations

Truncated seminorm spaces exhibit rich algebraic and topological structure:

• Intermediate Position: For compatible seminormed couples $(X_0,X_1)$, truncated interpolation spaces $(X_0,X_1)_{s,q}^\sigma$ are intermediate between the intersection and sum:

$$\Delta(X_0,X_1) \hookrightarrow (X_0,X_1)_{s,q}^\sigma \hookrightarrow \Sigma(X_0,X_1)$$

• Sum Decomposition: Truncated seminormed spaces decompose as algebraic sums with full-scale spaces, e.g.,

$\widetilde{B}^{s,p}_q = B^{s,p}_q + \dot{W}^{1,p}$

with precise control of seminorms via projection-operator constructions (Stevenson et al., 2020).

• Frequency-Space Characterization: Decomposition into low- and high-frequency parts (e.g., via Littlewood–Paley theory) reveals that truncation transitions a function space between Sobolev-type behavior at low frequencies and Besov-type at high frequencies (Stevenson et al., 2020).

7. Analytical and Probabilistic Implications

Truncated seminorms are intimately linked to both analytic and probabilistic constructs:

• Dirichlet Forms and Jump Processes: Truncated seminorms define energy forms for censored or reflected Lévy/Hunt processes, whose state-space evolution is restricted to neighborhoods determined by truncation (Rutkowski, 2018).
• Screened Scales for Trace and Embeddings: The use of truncation via the real interpolation method yields screened scales robust for traces, extension, and embedding theorems, including in unbounded domains and nonlocal frameworks (Stevenson et al., 2020).
• Coercivity in Nonlocal Models: Weighted truncated seminorms ensure coercivity and regularity results in PDEs with singular or degenerate coefficients associated to distance-to-boundary (Acosta et al., 13 Jan 2026).

A plausible implication is that truncated seminorms will continue to underpin further advances in nonlocal analysis, particularly in the study of function spaces adapted to irregular domains and in the probabilistic analysis of discontinuous Markov processes.