Instruments of direct democracy are often understood to be synonymous with “the sovereign will of the people” that is distinct and separate from representation. Their introduction into a constitutional democracy is then either invoked or disputed as an extraordinary device that can give a voice to “the people” or possibly surrender the powers of the government to “demagogues”. In this perspective, direct democracy is distinct from representation, can exist only occasionally and temporarily and necessarily opposes the processes of representation by a direct “presence” of the people, whether for good or for ill. Its power to provoke a “decision” of a popular sovereign is a possible source of deliverance in an age of disillusionment with elected officials or a possible threat to the tenets of constitutionally ordered and dominantly representative democracies. There is much truth to this perspective. Nonetheless, its uncritical reception in the quarters of constitutional theory resulted in an oversimplification. The “presence” of the people supposedly enabled by direct democracy is postulated to be neatly separated from representation in its structure. In this thesis, I demonstrate that this is not the case. Direct democracy involves distinct forms of representation. These feed off the basic feature of all procedures related to direct democracy, i.e. the effort to establish them as a lack of representation. A direct decision of the people is necessary because its interests are not or cannot be represented. By attempting to empty a particular space of representation, direct democracy becomes embroiled in representation as it has to articulate what representation (presumably) is not, what it is and what it may become. The constitutional order facing an act of direct democracy must then mediate between these different visions of representation rather than contending with a monolithic decision of “the people”. In performing its role of a mediator, a constitution that is to be applied to an act of direct democracy is itself stretched between structuring political processes and serving as their expression. Against such a background, a constitutional theorist should examine the different visions of representation and their interaction in the forge of direct democracy, rather than being “for” or “against” the awakening of the latter. In this dissertation, I investigate these claims by examining citizens’ initiatives and tracing their interaction with representation. Citizens’ initiatives are a form of direct democracy that allows the electorate itself to submit changes to existing legislative or even constitutional frameworks and possibly vote these into existence through referendums. They may thus be seen as an extreme form of “direct democracy”, where representatives are not even authorised to initiate the procedure and are in some cases excluded from deciding its fate. I demonstrate, however, that even such a radical innovation cannot credibly generate “direct democracy” that would be free from enacting representation. Referendums, another popular form of direct democracy, are placed in the context of citizens’ initiatives and their capacity to depart from representation is examined in this context. Three developments make this topic salient. Firstly, citizens’ initiatives are a part of Croatian constitutional law and scholars regularly take them to be a form of “direct democracy”, assuming its distinct nature that opposes it to representation. Secondly, contemporary constitutionalism is challenged by a number of contemporary populist movements that attempt to fix a particular understanding of representation through law. Given these developments, citizens’ initiatives become a particularly valuable way one may study how contemporary constitutions structure representation and how the law can be used to challenge and alter these structures. Thirdly, the global increase in the practice of direct democracy confronts contemporary constitutionalism with the need to facilitate a more profound understanding of what is at stake when “direct rule of the people” is invoked in a dominantly representative environment. Throughout this thesis, I have combined constitutional theory, political theory and comparative constitutional law to support my claims. As citizens’ initiatives are comparatively rare, it would not be sufficient to study them within the confines of a single legal order. Instead, I have turned to a comparative overview of the procedure in American, German, Italian and Croatian law. In so doing, I have primarily focused my attention on those states that have incorporated direct democracy and, unlike Switzerland, are self-described as dominantly representative democracies in which direct decision-making can make only an exceptional appearance. Swiss examples, by contrast, are used only exceptionally, as in this jurisdiction the representative process for the most part integrates direct democracy in an unprecedented degree. Besides this comparative approach, I have also had to develop a specific theoretical framework that would allow us to examine how the law structures and disputes representation. In order to achieve this aim, I have deployed constitutional and political theory in order to differentiate between three forms of representation: representation as an argument, process and construct. The law plays a role in making all three possible and in regulating their consequences for the constitutional framework that encompasses them. I examine the implications of this view throughout the dissertation. This thesis is divided into two parts. In the first part, I examine the concept of citizens’ initiatives and define them as a conflict centred on representation. The first chapter I demonstrate why citizens’ initiatives should be seen as a dispute in the first place. I discuss the different interpretations of “direct democracy” that can reflect on citizens’ initiatives, the vague subject of citizens’ initiatives and the indeterminate nature of the constitution, multiple relationships encompassed by the initiative, the lack of any specific content intrinsic to “direct decision-making” and the multiple functions that the initiative may fulfil. All these features may house more or less pronounced conflicts over representation, rather than being the realization of the power of the sovereign people. Having said this, however, citizens’ initiatives are not completely empty of content. They are interlaced with representation. To begin with, their introduction was motivated by problems identified in regular representative processes, which I discuss in the second chapter of the thesis. In addition, citizens’ initiatives are focus on the powers of the representative body and may affect its usual operation. While all this binds citizens’ initiatives to representation, it is common to understand them as being dedicated to the realization of “direct democracy”. The third chapter demonstrates that this distinction is a chimaera. In it, I show how it is impossible to assume that “the people” may be represented in one definitive act without annulling democracy and that the meaning of representation itself is not fixed. As this is the case, the initiative can only project a specific representation of the people and may articularate different senses of representation. As a result, a citizens’ initiative encompasses representation as an argument, representation as a process and representation as a construct. In order for the people to “directly decide”, the law enables these three different forms of representation and introduces their interaction. The second part of the dissertation examines the implications of the first part in the way citizens’ initiatives are structured and function. While citizens’ initiatives are normally seen as a linear procedure that starts with registration and ends with a decision, as is shown in the fourth chapter, the initiative as a conflict is composed of three sites of contestation: the subject of the initiative, its subject matter and the criteria used to determine whether the decision was actually reached by the people. Each of these three sites is examined in the three chapters that follow. The fifth chapter examines how the law is used in a twofold role within the initiative, to both express a political preference and contain the legitimate range of alternatives. I also demonstrate how this incorporates the clash between representation as an argument and representation as a process. The sixth chapter then expands on the topic of the subject of citizens’ initiatives. I explain the different methods used to articulate the “will of the people” and show how the initiative can only pose the question of who is being represented in the procedure, but cannot definitely settle it. This is further confirmed by the seventh chapter, in which I argue the initiative does not only incorporate the different methods of representing “the people” but also includes several different visions of the popular collective. The people appear simultaneously as a “group of individuals”, a “popular majority” and a “reasonable sovereign”. The final, eighth chapter is dedicated to citizens’ initiatives that contain a constitutive representative effect. These are initiatives that, besides encapsulating a dispute as to what it means to to represent the “will of the people” aim to substantially alter the representative process as it currently stands. While other initiatives may normally be absorbed by regular representative structures, constitutitive initiatives aim to amend these and thus pose a particular challenge. In their case, the dispute on representation extends beyond the boundaries of a particular initiative and may permanently reshape the sites where “the people” may make an appearance. In the last chapter of the thesis, I show how initiatives may influence the scope and structure of representative process and what are some of the constitutional boundaries to their effect in this regard. In conclusion, I show that, while constituent initiatives are not per se unconstitutional, they may become unacceptable insofar as, instead of being just one locus of the conflict on representation, they absorb other sites of conflict and subvert them to a particular vision of the “sovereign people”. Constitutional adjudication is in itself not necessarily sufficient to control such outcomes.