The notion of permission is of crucial importance in several settings, from law and ethics to artificial intelligence and normative reasoning. While deontic logic approaches often interdefine permission with obligation and prohibition, it is a complex notion with multiple interpretations. In this talk, I analyze permissions through the lens of preference logic, identifying and exploring three nuances: 'rather-so' permissions, advocating specific actions; 'better-not' permissions, allowing actions while expressing a preference against them; and neutral permissions, permitting actions without bias. The permissions are formalized via Hilbert axioms and ceteris paribus preferences and integrated into Standard Deontic Logic. By doing so, we create a richer framework that better reflects real-world decision-making, particularly in the fields of AI and law, where balancing permissions and preferences is crucial. We demonstrate how this framework can be applied to model complex decisions, offering a more nuanced approach to normative reasoning in these domains.