The question of the character of constitutional rights norms is complex and admits of no easy answer. Without reducing the complexity of the issue, I attempt in this paper to formulate some clear views on the matter. I shall argue that constitutional rights reasoning is a species of rational practical reasoning that combines both balancing (as Robert Alexy admits) and the grounds as to why balancing is appropriate (deontological constraints). Absent the latter type of reason, the application of constitutional principles remains a pure instance of balancing. Each time those reasons are touched upon, however, balancing becomes subject to deontological constraints. Yet deontological reasons are neither self-proclaiming nor complete. Rather they require for their articulation the background of an already operational practice of practical (legal) reasoning. This practice exemplifies what I shall dub the structure of autonomy, that is, a set of regulative ideals, not yet definitive norms, that derive from the reflective character of human agency. This structure comes to light when one turns to the agent’s point of view – to the point of view of someone who is engaged reflectively with practical questions (questions about rights are par excellence questions of this type). The structure of autonomy is more fundamental than any deontological or teleological reasons: it is in its light that deontology and teleology need to be understood as representing merely complementary forms of argument, which can be employed in reasoning that aims to maximise constitutional rights provisions. At the same time, the structure of autonomy as a regulative ideal generates prima facie reasons for allocating greater weight to deontological arguments in the relevant contexts of balancing. Such prima-facie reasons do not prescribe fully-fledged deontological constraints, but merely set the burden of proof in favour of specific reasons and ensure that the structure of autonomy be respected overall.