This thesis presents an attempt to understand understanding and explanation via the semantic and pragmatic study of 'why'-questions and their answers with 'because'. Despite their interdependencies, there is no single approach so far that examines both within one framework. The Relevance Nomological Model constructed for this purpose demonstrate (i) that speakers' principle-based knowledge (laws, rules, regularities of all sorts) is the crucial common component for the analysis of both 'why's and 'because's and (ii) how this knowledge intervene in the calculation of the felicity and truth conditions of 'why'-questions and 'because'-sentences, respectively. An answer to a 'why'-question is an instantiation of one or more abnormic clauses of a law and laws contained in speakers' background are the necessary ingredient for the analysis of 'because' and also of many other connectives. This hypothesis gave rise to the view of connectives as pragmatically tuned reasonings, whose premises contain (often) hidden law-like statements.