Recent alleged abuses of the pardon power by presidents in Peru and the US have drawn attention to a concept of the modern constitutional state that has largely operated under the radar: mercy. In addition to manifesting itself in the pardon power, mercy can also be found in the practices of judges who sentence defendants leniently. Both executive and judicial mercy have been criticised for conflicting with the rule of law and legal justice, which are key principles of the constitutional state. There are two main criticisms. First, pardons are usually understood as being unfettered by legal principles such as the principle of treating like cases alike, which is seen as conflicting with the rule of law. Second, by allowing judges to treat defendants more leniently than what legal justice requires, mercy is seen as being out of place in the constitutional state. What these criticisms have overlooked is that pardons and judicial mercy operate according to different logics. I argue in this thesis that they are modern-day instances of two different archetypes of mercy. Pardons are best conceived as an instance of the ancient Roman concept clementia, and judicial mercy is best conceived of as an instance of the Christian concept of misericordia. Deployments of both forms of mercy result in the lenient treatment of the beneficiary. However, they differ in that exercises of clementia do not require a specific motivation and are not constrained by law, which means that, legally speaking, the benefactor of clementia enjoys full discretion in either bestowing it or not. This is in contrast to exercises of misericordia that must be exercised for reasons of compassion and its exercise goes against the law. As I demonstrate, viewing executive and judicial mercy as instances of clementia and misericordia, respectively, is not only of historical interest. It also helps us understand why some of the criticisms levied against contemporary instances of executive and judicial mercy fail. Judicial mercy conflicts with legal justice and the rule of law due to its legacy of misericordia. But executive mercy, as an instance of clementia, does not. While not taking a stand on the respective desirability of these two forms of mercy, my thesis lays the foundation for informed future assessments of what, if any, role mercy should have in the constitutional state.