Modeling methods have been traditionally used for many aspects in legal informatics. Several of them feature visual notations to enable the understanding of complex relationships and ease human interaction with the corresponding models. Some also put a focus on the interpretation by machines and according processing functionalities. In this paper we regard abstraction and transparency as two functions in meta modeling, i.e. in the conceptualization of modeling methods. This permits us to highlight the nature of these two aspects and thus provide a basis for discussing the specific concepts of extraction and obfuscation, which play an important role for processing information in multi-sensory law and from a legal perspective in general. In this way applications for legal informatics can be analyzed and designed in regard to their abstraction and transparency capabilities.