For elaboration, please refer to the work. Kyburg discusses the proposition from (62).
Please note that this counterexample only works against a claim of always having a logical relation. A consideration of whether we are right most of the time is beyond (continued) this paper, but suffice it to say that humans are exponentially more accurate in relational judgments as the size of the discrepancy grows larger than our error "bounds."