Lower Bounds for Bit Pigeonhole Principles in Bounded-Depth Resolution over Parities
arXiv:2511.20023
Abstract
We prove lower bounds for proofs of the bit pigeonhole principle (BPHP) and its generalizations in bounded-depth resolution over parities (Res). For weak BPHP with pigeons (for any constant ) and holes, for all , we prove that any depth proof in Res must have exponential size, where is the number of variables. Inspired by recent work in TFNP on multicollision-finding, we consider a generalization of the bit pigeonhole principle, denoted -BPHP, asserting that there is a map from to () such that each has fewer than preimages. We prove that any depth proof in Res of -BPHP (for any constant ) must have exponential size. For the usual bit pigeonhole principle, we show that any depth Res proof of BPHP must have exponential size. As a byproduct of our proof, we obtain that any randomized parity decision tree for the collision-finding problem with pigeons and holes must have depth , which matches the upper bound coming from a deterministic decision tree. We also prove a lifting theorem for bounded-depth Res with a constant size gadget which lifts from -DT-hardness, recently defined by Bhattacharya and Chattopadhyay. By combining our lifting theorem with the -DT-hardness of the -variate Tseitin contradiction over a suitable expander, proved by Bhattacharya and Chattopadhyay, we obtain an -variate constant-width unsatisfiable CNF formula with clauses for which any depth Res proof requires size . Previously no superpolynomial lower bounds were known for Res proofs when the depth is superlinear in the size of the formula.
An earlier version containing one of the results appeared as ECCC TR25-118