The performance of an ontology alignment technique largely depends on the amount of information that can be leveraged for the alignment task. On the Semantic Web, end-users may explicitly or implicitly generate ontology alignments during their use of the semantic data. This kind of end-user-generated ontology alignment, which we call grass-roots ontology alignment, is an important source of information that is yet to be taken into account by current ontology alignment techniques. Grass-roots ontology alignment, often generated as a side effect of other data manipulations, could be user-specific, task-specific, approximate, or even contradictory. This paper reports our work on reusing grass-roots class alignment for aligning class hierarchies. A grass-roots class alignment, though approximate, still reveals some facts about relationships between different classes. We formalize facts about class relationships that can be inferred from an alignment under different cases. We then apply forward-chaining inference to the facts knowledge base to infer more facts. The facts KB is then leveraged for ontology alignment purposes. To deal with uncertainty and inconsistency, each fact is associated with an evidence that tells how the fact is obtained. The evidences are used to select better-supported facts in case of inconsistency.