Is it true that in most peer-reviewed journals the editorial office can decide to publish a paper without at least one reviewer approving some version to be published as is? I am asking whether they technically have the right to do so in some kind of exceptional situation. Say, for instance, the reviewer has approved the paper for publication modulo some minor corrections, but after these corrections have been made they stop responding to communication.
Or do most journals have strict protocols forbidding publication without a final approval from at least one reviewer? In case "most peer-reviewed journals" is too broad, my field is pure mathematics.