City Research Online

Abstract

This thesis explores the drivers of Mutual Adjustment in digital organizations, a process through which agents influence each other's adaptation processes. While existing research has extensively examined Mutual Adjustment in traditional organizational settings, novel forms of organizing challenge the boundary conditions of established theories. Hence, this study leverages naturally occurring data from open-source software communities and decentralized autonomous organizations to investigate three facets of Mutual Adjustment: coordination, authority, and control. By examining organizations where actors operate without formal employment relationships, this research enquires about the boundaries and nature of Mutual Adjustment and the scope of applicability of extant managerial knowledge.

Chapter 2 examines how organizations rely on Mutual Adjustment after adopting a Plan. Using a 20-year longitudinal dataset from the Linux Kernel project, it analyzes the introduction of a plan in 2005. Findings suggest that planning allows for widening the scope of Mutual Adjustment but with lower needs per action awaiting coordination. Conversely, when the plan fails, cognitive resources are depleted by widening the Mutual Adjustment scope and increasing the needs per action. The study further shows that agents redistribute freed resources based on their representation of the network of interdependence.

Chapter 3 explores the emergence of legitimate authority in 'supposedly' flat organizations. In particular, the chapter investigates what gives individuals the legitimate authority to influence others and the work they do, problematizing the role of formal structure. By leveraging a longitudinal dataset of 10 DAOs, the chapter finds that formal leader-subordinate relationships persist even in decentralized settings. However, formal leaders face a penalty compared to comparable alters, suggesting that structure alone is not consequential for legitimate authority.

Chapter 4 investigates how natural language is used as a control device for technology decisions. Using naturally occurring data from the Linux Kernel project, it integrates a qualitative analysis of meanings with a Machine Learning text classifier to identify linguistic drivers of control. The chapter proposes control attempts emerging in the non-random association between linguistic drivers and four specific objectives of control – i.e., external boundaries, internal boundaries, style, and expected utility. The chapter tests the emerging control attempts against the scope of participation and contribution. Overall, the chapter findings highlight how linguistic drivers enable organizational agents to find footing with each other.

As a reflection on the method applied in Chapter 4, the final chapter4 discusses potential challenges and solutions to qualitatively analyze a large corpus of textual data. In so doing, it presents a methodological framework that combines traditional qualitative methods with supervised and unsupervised Machine Learning applications to enhance and scale human-led analysis. The framework leverages Machine Learning mapping and scaling capabilities to orient human attention and energy toward substantive semantic instances, facilitating the discovery, refinement, and evaluation of emerging organizational insights.

Keywords: Mutual Adjustment, Language, Coordination, Authority, Control, Plan, Formal Structure, Epistemic Interdependence, Machine Learning, Open-Source Software, Decentralized Autonomous Organizations.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HD Industries. Land use. Labor > HD28 Management. Industrial Management
Departments:	Bayes Business School > Bayes Business School Doctoral Theses Bayes Business School > Faculty of Management Doctoral Theses

Export

Downloads