We define a new NP search problem, the “local improvement” principle, about labellings of an acyclic, bounded-degree graph. We show that, provably in , it characterizes the consequences of and that natural restrictions of it characterize the consequences of and of the bounded arithmetic hierarchy. We also show that over V0 it characterizes the consequences of V1 and hence that, in some sense, a miniaturized version of the principle gives a new characterization of the consequences of . Throughout our search (...) problems are “type-2” NP search problems, which take second-order objects as parameters. (shrink)

Conservative subtheories of $${{R}^{1}_{2}}$$ and $${{S}^{1}_{2}}$$ are presented. For $${{S}^{1}_{2}}$$, a slight tightening of Jeřábek’s result (Math Logic Q 52(6):613–624, 2006) that $${T^{0}_{2} \preceq_{\forall \Sigma^{b}_{1}}S^{1}_{2}}$$ is presented: It is shown that $${T^{0}_{2}}$$ can be axiomatised as BASIC together with induction on sharply bounded formulas of one alternation. Within this $${\forall\Sigma^{b}_{1}}$$ -theory, we define a $${\forall\Sigma^{b}_{0}}$$ -theory, $${T^{-1}_{2}}$$, for the $${\forall\Sigma^{b}_{0}}$$ -consequences of $${S^{1}_{2}}$$. We show $${T^{-1}_{2}}$$ is weak by showing it cannot $${\Sigma^{b}_{0}}$$ -define division by 3. We then consider what (...) would be the analogous $${\forall\hat\Sigma^{b}_{1}}$$ -conservative subtheory of $${R^{1}_{2}}$$ based on Pollett (Ann Pure Appl Logic 100:189–245, 1999. It is shown that this theory, $${{T}^{0,\left\{2^{(||\dot{id}||)}\right\}}_{2}}$$, also cannot $${\Sigma^{b}_{0}}$$ -define division by 3. On the other hand, we show that $${{S}^{0}_{2}+open_{\{||id||\}}}$$ -COMP is a $${\forall\hat\Sigma^{b}_{1}}$$ -conservative subtheory of $${R^{1}_{2}}$$. Finally, we give a refinement of Johannsen and Pollett (Logic Colloquium’ 98, 262–279, 2000) and show that $${\hat{C}^{0}_{2}}$$ is $${\forall\hat\Sigma^{b}_{1}}$$ -conservative over a theory based on open cl-comprehension. (shrink)

The complexity class of [Formula: see text]-polynomial local search problems is introduced and is used to give new witnessing theorems for fragments of bounded arithmetic. For 1 ≤ i ≤ k + 1, the [Formula: see text]-definable functions of [Formula: see text] are characterized in terms of [Formula: see text]-PLS problems. These [Formula: see text]-PLS problems can be defined in a weak base theory such as [Formula: see text], and proved to be total in [Formula: see text]. Furthermore, the [Formula: (...) see text]-PLS definitions can be skolemized with simple polynomial time functions, and the witnessing theorem itself can be formalized, and skolemized, in a weak base theory. We introduce a new [Formula: see text]-principle that is conjectured to separate [Formula: see text] and [Formula: see text]. (shrink)