Ontologies are utilized for a wide range of tasks, like information retrieval/extraction or text generation, and in a multitude of domains, such as biology, medicine or business and commerce. To be actually usable in such real-world scenarios, ontologies usually have to encompass a large number of factual statements. However, with increasing size, it becomes very difficult to ensure their complete correctness. This is particularly true in the case when an ontology is not hand-crafted but constructed (semi)automatically through text mining, for example. As a consequence, when inference mechanisms are applied on these ontologies, even minimal inconsistencies oftentimes lead to serious errors and are hard to trace back and find. This paper addresses this issue and describes a method to validate ontologies using an automatic theorem prover and MultiNet axioms. This logic-based approach allows to detect many inconsistencies, which are difficult or even impossible to identify through statistical methods or by manual investigation in reasonable time. To make this approach accessible for ontology developers, a graphical user interface is provided that highlights erroneous axioms directly in the ontology for quicker fixing.