After many year the phd is finally submitted (last November) and the defense is planed for February. Let me give the abstract and executive summery.
This PhD thesis is an interdisciplinary investigation of how the creation of novelty can be regulated. Novelty is a general concept describing the evolutionary emergence of things that did not exist before, while integrating more specific ideas like adaptation, exaptation, creativity, invention and innovation. The controlled production of novelty must construct radically new things out of existing constraints, and thus overcome the conservation law that is inherent in most phenomena. A general solution to this paradox is proposed in the form of an abstract novelty model. It shows how novelty can bootstrap itself by means of a network of feedback loops interacting via a central workspace that mediates between system and environment. This abstract model is then mapped onto four more concrete models, showing how novelty emerges in four different domains: individual cognition, evolution of science and technology, strategic management of organizations, and management of socio-technological innovation. A meta-model then shows how these different applications can be mapped onto each other through a grid of conversion patterns. This suggests two methodologies for implementing the new insights: Carbureted Action Research, which proposes a development path for the phase transition that characterizes the emergence of a socio-technological ecosystem; and the Agile-Enterprise Innovation Planning architecture, which makes such radical innovation more manageable by stimulating the spinning-off of supporting ecosystems. To validate these methodologies, concrete applications in three domains are investigated: an experiment with project-based educational systems, a case study of an open-source software development ecosystem, and a design for an “Interversity”, i.e. Internet-based university of the future. The resulting observations provide a proof of concept that the proposed novelty theory is applicable in practice.
This PhD thesis is an interdisciplinary investigation, inspired by evolutionary and cybernetic theories, of how the creation of novelty can be regulated. Novelty is what did not exist before, but emerges into existence. Novelty is an abstract concept that binds together more concrete ideas like creativity, discovery, invention and innovation, while clarifying the differences between natural and artificial evolution.
Natural evolution produces relative novelty through adaptation. Artificial evolution produces absolute novelty through exaptation, i.e. the utilization of a structure for a function other than that for which it was developed through environmental selection. Technology is an example of absolute novelty, but so are proteins and feelings. In contrast to the relative novelty that evolves gradually through adaptation, radical novelty is the result of a phase transition in which a new type of system emerges. Radical and absolute novelties cannot be explained by mere gradual adaptation: they require a mechanism of novelty production.
Models of such a mechanism have been proposed in different domains, including cognition, science and technology studies, and strategic management of innovation. The present work generalizes and synthesizes several of these models in the form of an abstract novelty model. This abstract model is then mapped onto more concrete models in six different application domains: sailing against the wind, knowledge creation, environmental enrichment, strategy management, agile development, and systematic innovation.
The essence of the abstract model is that it shows how the process can overcome the conservation law that is inherent in most phenomena, and which basically states that something cannot be created out of nothing. For example, sailing against the wind seems to contradict the law of conservation of momentum, according to which you cannot convert momentum moving in one direction (the one in which the wind blows) to momentum for moving in the opposite direction (the one to which the ship is heading). In the case of knowledge creation, the novelty producing process must overcome the conservative bias inherent in the fact that all problem solving is based on existing knowledge. In the case of strategic management, the existing organization imposes constraints that make it difficult to create a truly novel organization.
Thus, the controlled production of novelty is inherently paradoxical, as it must use existing resources to create radically new ones. The novelty model proposes a general solution to this conundrum by showing how novelty can bootstrap itself out of existing elements by means of a system of feedback loops interacting via a workspace.
This theoretical model is then applied as a conceptual tool to support organizational and technological innovation. It does so by proposing two general methodologies: Agile-Enterprise Innovation Planning (ÆIP) and Carbureted Action Research (CAR). To validate these methodologies, concrete applications in three domains are investigated: an experiment with project-based educational systems, a case study of an open-source software development ecosystem, and a design for an “Interversity”, i.e. Internet-based university of the future. With these different validations, this PhD thesis provides a proof of concept that the proposed novelty theory and its methodology to support systematic innovation are applicable to the real world.
After a general introduction to the problem domain, the thesis begins with a chapter that uses an agent-based simulation and a number of illustrative examples of technologically supported cognition to examine how intelligence can be amplified. We thus investigate the mechanisms behind exaptation and the serendipitous discovery of new functions. The research draws our attention to how important the enrichment of the environment is for augmenting the intelligence of agents acting in that environment.
A deeper understanding of the mechanism for intelligence amplification is required, motivating the development of the novelty theory. From the study of cognition we move to fundamental studies of complex evolution, including the self-organization of dissipative structures and the origin of life. This helps us to understand how the evolutionary mechanism overcomes conservation—i.e. how it produces novelty—by using a workspace. By default, the workspace is the medium for agents to work in. By giving the concept "workspace" a more general meaning, we can use it to better understand how environmental conditions can amplify intelligence.
In the next chapter, the general novelty model is introduced as a bootstrapping mechanism on a workspace that can produce novelty more effectively. The workspace is subjected to four processes, called "novelty anchors", that feed in and out of it: internalizing, externalizing, directing and evolving. The workspace is the central medium of operation. It is connected via the anchors to two other mediums: the environment (external) and the system (internal). The artificial evolution of the workspace develops both the system core (e.g. internal memory) and the environmental enrichment. This produces a constructive co-evolution between system and environment.
To make the abstract novelty model more concrete, we apply it to the case of individual cognition. Here, we show how knowledge is created in a coherent way, which is why this application is called the Cohering model. Modeling knowledge and mastering action describe the two aspects of the co-evolution process between system and environment. Modeling is the process that develops the system core on the basis of the novelty (i.e. knowledge) that emerges in the workspace. Mastering is the complementary process that implements the novelty by enriching the environment. Advanced cases of modeling and mastering bring us to the need for studying scientific and technological development.
From Science and Technology Studies (STS) we learn how constructs in the social fabric can mobilize elements of the world, thus bringing together resources that were initially distributed around the globe into one socio-technological ecosystem. This makes these resources more accessible for study and use. In his study of mobilization, the STS scholar Bruno Latour has analyzed the novelty production more explicitly. This inspires us to formulate a second concrete implementation of the abstract novelty model, which maps onto Latour’s model of mobilization. This new model, the Eventuating model, describes the modeling of workspaces and the mastering of scarcity. By means of historical examples, it is shown how phase transitions in socio-technological ecosystems can revolutionize the whole social fabric.
The phase transition characterizing innovation becomes a central issue in Strategy Management Studies (SMS). First, we consider another application of the novelty model, called the Strategizing model, which examines the value that organizations create. Studies of organizations in SMS have led to the development of a framework that describes the dynamic capabilities of an organization. The dynamic capabilities framework becomes another application of the novelty model, called the Strategizing model. It investigates how modeling and mastering apply to organizations. Different types of innovation are situated in the different stages of the phase transition, thus giving us a broader view of the process of innovation.
Innovation requires project management and methods for designing solutions in a complex adaptive environment. This brings us to another application of the novelty model, called the Establishing model, which helps us to produce effective project management for innovation. With this last application we are ready to formulate a meta-model, which integrates the four previous models (Cohering, Eventuating, Strategizing and Establishing). This meta-model allows us to investigate the phase transition more systematically, by focusing on conversion patterns (e.g. the conversion from implicit knowledge to explicit knowledge). Each of the novelty models has its own conversion patterns. Together, these conversion patterns define the Agile-Enterprise Innovation Planning (ÆIP) grid. The ÆIP grid is a form of descriptive knowledge, i.e. it explains how innovation happens. For practical applications, we also need prescriptive knowledge that would tell us how to produce innovation.
The prescriptive model, which is called Carbureted Action Research (CAR), proposes two co-development paths that envelop the innovation during the phase transition. The final part of the CAR model is the ÆIP architecture, which is an Enterprise Architecture that shows how supporting socio-technological ecosystems can be spun off from the main innovation. The ÆIP architecture is intended to make radical innovation more manageable. To make the ÆIP architecture more concrete, three different validations are elaborated. The first validation uses controlled experiments in education. The second one is based on participation research in an open-source community. The last one is more theoretical, proposing an operational model for the ÆIP architecture in the context of universities.
The controlled experiments focus on project-oriented education about software tools. In the agile education experiment, we investigate how to teach when the subject is evolving so quickly that the content is outdated by the end of the education period—a phenomenon characteristic of Internet-based software development. The solution is to focus more on the methods than on the content and to shift from education as knowledge transfer from teacher to student to education as teacher-supervised knowledge creation by the students. However, the bottleneck for agile education is the need for close supervision, which requires a low student-to-teacher ratio.
In the scalable education experiment, complementing teacher supervision with peer evaluation widens this bottleneck. This in principle allows the education system to scale up to a much larger number of students for the same number of teachers. Evaluation of students’ work by other students requires us to develop a framework of methods and software tools to ensure a sufficient quality of results (e.g. by motivating students to make a fair number of contributions and by minimizing the effect of poor or biased evaluations).
In these teaching experiments, students are asked to develop a prototype application of a software tool for doing Internet-supported business—ideally in order to create a spin-off company. This shows how, in the ÆIP architecture, education can be an interface between academic research and a concrete development ecosystem outside the university.
By participating in an actual software development project, namely the Drupal open-source content management system, we obtain a deeper insight in the concrete dynamics of self-organizing innovation. We observe an interesting business ecosystem emerging through a variety of capital investments. By interviewing the entrepreneurs in the ecosystem, we get a more detailed view of how innovation happens in practice. This provides us with a better understanding of the natural development of an innovative ecosystem, which we need for the ÆIP architecture.
The last challenge for the ÆIP architecture is to create an operational model that combines the insights of the other validations to design a system that can spin off innovation ecosystems. The envisaged system is a self-organizing and distributed university, called the Interversity, where project-oriented, research-based education is central. The focus is on the development of ecosystems that support the research within the university, but also provide a wider service to society. As such, the Interversity would rest on the same three pillars as the traditional university (research, education and public service), but with much stronger synergy between these pillars. The goal of the operational model is to provide a proof of concept for the ÆIP architecture, and in particular propose a direction for future research on the subject.
The concluding chapter summarizes the main principles of the novelty theory developed in this dissertation, together with its potential applications in supporting innovation. It is followed by an extensive glossary in which the most important new concepts discussed in this work are briefly defined.