As reported in the literature, one of the key problems of conceptual modeling in critical domains is the cognitive intractability of the resulting models. Given the typical intrinsic complexity of these domains and the expressivity required by the models so that they can really serve as precise characterizations of the domains at hand, these models tend to rapidly grow in size and in the level of detail to contain thousands of classes, relations, generalization partitions, etc. Our objective is to build on existing literature on clustering, relevance, summarization and filtering methods. In our approach instead, by relying on a conceptual modeling language that contains rich ontological foundation, we can leverage on natural domain-independent mechanism for viewpoint selection, clustering, relevance determination and abstraction. In the sequel, we elaborate on four complementary mechanisms for cognitive complexity management of large conceptual models.