In this thesis, I challenge assumptions underlying my discipline of enterprise architecture that led to two choices facing practitioners: either to work with tools and techniques which predict and control changes towards predetermined ends or to accept informal processes that are unpredictable and wasteful. Orthodox enterprise architecture defines an enterprise as an organisation, which is a system, and prescribes methods that seek to provide control over the transformation of an organisation into a desired state of affairs by achieving complete knowledge of the system before initiating the desired transformation. Drawing on complexity sciences, I offer a different perspective on organisation and claim that organising what we do is an aspect of doing what we do. Organising is process. I furthermore claim that the people who are organising what we do can act spontaneously and surprise both themselves and others, but often they act habitually. Habitual ways of acting allow us to anticipate to some extent how others are likely to respond to us and, as we grow up, we learn how to behave ourselves, that is, how to adjust our behaviour to what we judge socially acceptable to increase the likelihood of being able to garner support and collaboration. I posit that social control is exercised in this way as mutual self-adjustment that forms what is normal and valued conduct. In other words, our shared social norms and values thus paradoxically and simultaneously form individuals and their conduct and are formed by individuals and their conduct. I claim that in this way we have partial, but never full, knowledge of how others generally respond to certain behaviour of ours. We can ever have only partial knowledge of that which is - in the words of Mannheim - in the process of becoming. I therefore reject the central assumptions upon which orthodox enterprise architecture is based. In organisations, we engineer and exploit mechanical mechanisms that can conduct certain action more effectively and efficiently than people can. Materiality, objects in the world, can resist attempts to shape them to suit our needs but do so without intentionality or spontaneity. Accommodating material resistance is thus repeatable. Enterprise architecture as a discipline grew out of engineering of physical mechanisms and assumes a similar repeatability and predictability when working with the social, which I find to be an unwarranted assumption. I argue against the claim of orthodox enterprise architecture that we can bring about a pre-determined state in a controlled fashion and against the claim that without such control we have informal processes that are inevitably unpredictable and wasteful. I posit that what emerges is paradoxically stable instabilities of socially enabled and constrained recognisable patterns of behaviour. When devising a mechanism in a physical object, such as a software programme, a repertoire of scripted action is transcribed into it which remains constant until transcription is renewed. Transcription has a tendency to render action less fluid. Some members of an organisation may judge particular scripted action to be awkward or detrimental while others may judge the same scripted action to be efficient and beneficial. Thus, determining which scripted action to transcribe into mechanisms is a highly political decision which attracts the attention of skilful political players. Enterprise architects can have a valuable role to play, since we have a better than average partial knowledge about technology, and since technology is increasingly important for many enterprises. I posit that becoming more aware of power and power plays, developing a feel for the game, and becoming more detached about our involvement will allow us to play into what is emerging socially with more political awareness and expertise.

Afin de réduire les pollutions agricoles, les agriculteurs sont invités à faire évoluer leurs pratiques. Mais ces évolutions génèrent des changements d'organisation et de temps de travail, qu'ils invoquent pour expliquer leur difficulté à changer. Notre objectif est d'étudier ces changements du travail, pour concevoir des outils facilitant les transitions agroécologiques des agriculteurs. Nous avons enquêtés des conseillers agricole et des agriculteurs pour analyser l'offre de conseil, et la nature des informations sur le travail, dont les agriculteurs ont besoin pour conduire un changement de pratiques. Constatant l'insuffisance des outils existants, nous avons organisé deux ateliers de conception d'outils avec leurs futurs utilisateurs. Nous avons identifié et décris 28 outils qui pourraient être conçu. Quatre prototypes ont été élaborés : une fiche technique informant des changements du travail, un quizz et un jeu permettant d'identifier des changements du travail pour une exploitation, et un carnet de suivi d'essai permettant de relever des informations clefs pour établir un bilan de la gestion des concurrences et des choix d'organisation. / In order to reduce agricultural pollution, farmers are invited to change their practices. But these new practices can generate changes in organization and working time, which farmers assert to explain their difficulty to change. Our goal is to study these changes in the work and to design tools facilitating agroecological transitions of farmers. We have surveyed agricultural advisers and farmers to analyze the consulting service offer, and the nature of the work information that farmers need in order to implement new practices. Noting the inadequacy of the existing tools, we organized two tool design workshops with their future users. We have identified and described 28 tools that could be designed. Four prototypes were developed: a technical sheet informing about work changes, a quiz and a serious game to identify work changes for a farm, and a test logbook to record key information to establish an assessment of work competition management and organizational choices.