The aim of this thesis is to describe and defend 'The Constraint-Based View', which is a particular conception of the nature of, and relationship between, ideal and non-ideal theories of justice. Traditionally, ideal theory is characterised by the assumption of 'full compliance', while non-ideal theory is characterised by the assumption of 'partial compliance'. In other words, ideal theories assume that those for whom the theory is meant to apply will be entirely willing and able to comply with the theory's requirements, while non-ideal theories do not.In Chapter 1 of this thesis, I describe and assess this original conception as well as several alternative accounts of ideal and non-ideal theory, in order to offer a broad survey of the existing literature, and to identify the various ways that these conceptions fail to capture fully the relationship between ideal and non-ideal theory.In Chapter 2, I draw a distinction between two different approaches to theorising about justice. 'Axiological' or 'A-Type' approaches are characterised by the fact that they include almost no assumptions as inputs to the theory, and that they are not intended to provide action-guiding recommendations as part of the outputs of the theory. In contrast, 'Practical' or 'P-Type' approaches include additional assumptions as part of their inputs and are intended to form part of a process that ultimately produces action-guiding recommendations. In Chapter 3, I describe and defend my preferred conception of the relationship between ideal and non-ideal theory - the 'Constraint-Based View' (CBV). According to the CBV, there is a spectrum of theoretical approaches that can be more or less ideal, depending on the extent to which they include 'soft constraints' as part of the inputs to the theory. Soft constraints are facts about the world that can be changed, in contrast to hard constraints, which are facts about the world that cannot. I argue that this way of thinking about the relationship between ideal and non-ideal theory is more useful than the alternative conceptions considered in Chapter 1. In Chapter 4, I examine the roles that feasibility considerations should play in ideal and non-ideal theory, from the perspective of the CBV and in Chapter 5, I explain the implications of adopting the CBV for what I call 'transitional theory', which is concerned with the transition from the status quo towards a more ideal state of affairs. In Chapter 6, I offer an example of the CBV in action, by considering its implications for the debate over duties of justice towards future generations.