This thesis studies knowledge processes which span organisational boundaries, examining how knowledge is formed and shared between two companies with divergent interests, facing the challenges of innovation processes. Cross-boundary work provides access to diverse sources of knowledge, specialties and approaches, and this enhances the innovative performance of firms. However, managing knowledge and spanning diverse boundaries has proven to be difficult. While the epistemic and social challenges have been identified to trigger conflicts and misunderstandings across boundaries, the complexity, inter-dependency and uncertainty of innovation processes have been found to multiply these challenges. Informed by the practice-based perspective, this thesis examines how the challenges of integrating knowledge in such a context are addressed and resolved. Building upon this theoretical perspective, the thesis aims to enhance understandings of knowledge processes between the large bureaucratic organisations (publishers) and small/medium-sized companies (developers) in the videogames industry. Underpinned by a social constructionist and interpretivist methodology, a qualitative study of three publisher-developer relationships was conducted. With thirty six semi-structured interviews with senior directors, managers and team leaders, the thesis examined cross- boundary practices and the conflicts experienced in this process. In order to achieve this, the study focused on the role that boundary objects, trust development and power structures played in facilitating knowledge processes. The thesis recognises the evolving and relational character of boundary objects, highlighting that a combination of static and dynamic boundary objects were effective in facilitating knowledge integration in the publisher-developer relationship. It also reveals that despite high levels of distrust between the parties involved, they managed to create and maintain a working relationship by resorting to opportunistic practices, such as knowledge hiding, deception and collusion. It is emphasised that understanding the power dynamics in the publisher-developer relationship is a pre-requisite to explaining 2 how knowledge is managed across organisational boundaries during the course of a project. Drawing upon a Foucauldian perspective, the thesis identifies the positive and enabling aspects of power dynamics in this relationship. It argues that when the parties have discrepancies, competition and high levels of distrust, power games positively influence cross-boundary practices, the use of boundary objects and knowledge processes, ultimately mobilising knowledge integration. The thesis makes four significant contributions to the knowledge management and cross-boundary work literature. First, it identifies an evolving role for boundary objects, showing how they develop during the course of a project. Second, it finds a relational and politicised dimension for boundary objects, highlighting the role of brokers to manipulate and mobilise the use and effectiveness of these objects. Third, the thesis extends the existing literature by revealing that despite high levels of distrust, parties can create a working relationship. The research shows that this is achieved through resorting to opportunistic behaviour, such as knowledge hiding, deception and collusion. As a result, this thesis adds a complementary level to Carlile's Integrated Framework, explaining that when there are high levels of differences, dependencies and novelty in the cross-boundary work, knowledge hiding, deception and collusion are the driving force to facilitate knowledge integration and maintain a functional relationship. The fourth contribution of this thesis is recognising the positive and productive aspects of power dynamics that enable and mobilise boundary objects and knowledge processes, ultimately bringing positive outcomes and creating a functional relationship between two companies with divergent interests.

Public rangelands are managed by a mixture of federal, state, and local governments. Often, these groups are charged with managing adjacent lands that are part of the same greater landscape. To do this effectively, communication and collaboration is required. This dissertation examines federal, state, and local agencies’ level of communication through three projects. The first project examined barriers to agencies adopting management tools from each other. I found that individuals within agencies were mainly staying within their own agency when seeking advice, so individuals were not communicating about tools or their findings across agencies. Furthermore, agency policies and fear of being sued restricted individuals’ ability to adopt management tools. The second project studied how land and wildlife managers in Southeastern Utah work together, or not, in managing mule deer populations that migrate to and from land managed by different agencies. I found that managers are working together to manage mule deer populations, but there are only a few individuals that tie everyone together. The third project looked at how federal, state, and local governments work together to rehabilitate lands after a wildfire that burned parts of Southwestern Idaho and Southeastern Oregon. I found that policy decisions at the federal level can heavily impact who works together and when. All three projects revealed that there are still barriers to federal, state, and local governments working together to manage the same landscape. However, the results from this dissertation also highlight opportunities for bridging the gap between agencies and, ultimately, improving management of rangelands.