Polynomial definability in constraint languages with few subpowers
Jakub Bulín, Michael Kompatscher
TL;DR
The paper investigates when constraint languages admit short pp-definitions and whether this property aligns with the classical few subpowers condition. It proves a main universal-algebraic result: for algebras with a $k$-edge term whose variety is residually finite, all pp-definable relations have length $O(n^{\max(2,k-1)})$, yielding short pp-definitions in many important cases, including all 3-element domains. The work further links short pp-definitions to the subpower membership problem (SMP), showing NP ∩ co-NP containment and highlighting certificates as a vital tool for potential P-time algorithms. It also discusses extensions beyond residual finiteness, interpretability considerations, and open questions about computability and degree bounds, offering a roadmap for strengthening tractability analyses in CSPs through universal algebraic techniques.
Abstract
A first-order formula is called primitive positive (pp) if it only admits the use of existential quantifiers and conjunction. Pp-formulas are a central concept in (fixed-template) constraint satisfaction since CSP($Γ$) can be viewed as the problem of deciding the primitive positive theory of $Γ$, and pp-definability captures gadget reductions between CSPs. An important class of tractable constraint languages $Γ$ is characterized by having few subpowers, that is, the number of $n$-ary relations pp-definable from $Γ$ is bounded by $2^{p(n)}$ for some polynomial $p(n)$. In this paper we study a restriction of this property, stating that every pp-definable relation is definable by a pp-formula of polynomial length. We conjecture that the existence of such short definitions is actually equivalent to $Γ$ having few subpowers, and verify this conjecture for a large subclass that, in particular, includes all constraint languages on three-element domains. We furthermore discuss how our conjecture imposes an upper complexity bound of co-NP on the subpower membership problem of algebras with few subpowers.
