• Nie Znaleziono Wyników

On the Potential to Manage a Transition to Sustainability in the Westland

N/A
N/A
Protected

Academic year: 2021

Share "On the Potential to Manage a Transition to Sustainability in the Westland"

Copied!
244
0
0

Pełen tekst

(1)
(2)
(3)
(4)

S

W

Proefschrift

ter verkrijging van de graad van doctor aan de Technische Universiteit Delft,

op gezag van de Rector Magnificus prof. ir. K.C.A.M. Luyben, voorzitter van het College voor Promoties,

in het openbaar te verdedigen op vrijdag 11 december 2015 om 12:30 uur.

door

J. K

ASMIRE

MSc. in the Evolution of Language and Cognition, geboren te Sacramento, California, USA.

(5)

promotor: prof.dr.ir. Gerard P.J. Dijkema copromotor: Dr. Ir. I. Nikolic

Samenstelling promotiecommissie:

Rector Magnificus, voorzitter

Prof.dr.ir. M.P.C. Weijnen, Technische Universiteit Delft Prof.dr.ir. Gerard P.J. Dijkema, Rijksuniversiteit Groningen Dr.ir. I. Nikolic Technische Universiteit Delft Onafhankelijke leden:

Prof. William Nuttall The Open University, UK Prof. Arthur C. Petersen, University College London, UK Prof. Eswaran Subrahmanian Carnegie Mellon University, USA Prof. Paulien Herder Technische Universiteit Delft

Prof. R.W. (Rolf ) Künneke Technische Universitiet Delft, reservelid

Keywords: complex adaptive systems, evolution, sustainability, transition man-agement, transitions, emergence, agent-based modelling

Printed by: Gildeprint Drukerijen

Front & Back: ‘Night Food: Decisions’ by Robert Turner https://500px.com/theoddnetwork email: jkasmire@googlemail.com

Copyright © 2015 by J. Kasmire ISBN 978-94-6233-134-1

(6)

accompanying the dissertation

O

N THE

P

OTENTIAL TO

M

ANAGE A

T

RANSITION TO

S

USTAINABILITY IN THE

W

ESTL AND

by

J. K

ASMIRE

1. The present collapses the adjacent possible future into the retreating actual past. 2. Scientific understanding co-evolves with approaches and theories.

3. Like facts (Arbesman,2012), theories have a half-life. 4. Equilibrium is an unnatural state for natural systems. 5. Mistakes are a useful and unpredictable form of creativity.

6. The anthropocene era will be recognised by the ubiquity of plastic in sedimentary layers and the word ‘sustainable’ in written records.

7. Agreement and certainty are impediments to intellectual evolution.

8. An emergent feature of car ownership is the spontaneous appearance of a ‘Best of...’ album that no one remembers buying.

9. The theoretical or practical value of an idea cannot be foreseen.

10. If life is the propagation of dimensions (Kauffman,2002), then the curse of dimen-sionality and the ergodic hypothesis allow living systems to reduce entropy.

These propositions are regarded as opposable and defendable, and have been approved as such by the promotor prof. dr. ir. M. Weijnen.

(7)

Stellingen

behorende bij het proefschrift

O

N THE

P

OTENTIAL TO

M

ANAGE A

T

RANSITION TO

S

USTAINABILITY IN THE

W

ESTL AND

door

J. K

ASMIRE

1. Het heden stort de aangrenzende mogelijke toekomst in het terugtrekkende wer-kelijke verleden.

2. Wetenschappelijke kennis co-evolueert met benaderingen en theorieën. 3. Net als feiten (Arbesman,2012), hebben theorieën een halfwaardetijd. 4. Evenwicht is een onnatuurlijke toestand voor natuurlijke systemen. 5. Fouten zijn een nuttige en onvoorspelbare vorm van creativiteit.

6. Het Antropoceen zal herkend worden aan de alomtegenwoordigheid van plastic in sedimentaire afzettingen en het woord ‘duurzaam’ in schriftelijke documenten. 7. Overeenstemming en zekerheid zijn belemmeringen voor intellectuele evolutie. 8. Een emergent kenmerk van autobezit is de spontane verschijning van een ‘Best

of...’ album waarvan niemand zich herinnert het gekocht te hebben.

9. De theoretische en praktische waarde van een idee kan niet worden voorzien. 10. Als leven de uitbreiding van dimensies is (Kauffman,2002), dan maken de vloek

van dimensionaliteit en de ergodische hypothese het mogelijk dat levende syste-men entropie verminderen.

Deze stellingen worden opponeerbaar en verdedigbaar geacht en zijn als zodanig goedgekeurd door de promotor prof. dr. ir. M. Weijnen.

(8)

List of Figures vii

List of Tables ix

Acknowledgements xi

1 Introduction 1

1.1 Research . . . 3

1.2 Anticipated relevance, scientific contribution and audiences. . . 4

1.3 Reader’s Guide to this thesis. . . 5

I The greenport questions: Applying TM to the Westland-Oostland Greenport 7 2 Understanding Transition Management 9 2.1 Important underlying concepts. . . 10

2.1.1 CAS . . . 10

2.1.2 Transitions. . . 15

2.1.3 Sustainability . . . 21

2.1.4 Modern governance theory . . . 24

2.2 TM itself . . . 25

3 Historical analysis and case study 31 3.1 A case study of CHP units in the greenport . . . 32

3.1.1 Introduction, theoretical framework and methodology . . . 33

3.1.2 Case study results . . . 35

3.1.3 Case study discussion . . . 41

3.2 A historical analysis of greenhouse horticulture. . . 42

3.2.1 Introduction, theoretical framework and methodology . . . 43

3.2.2 Historical analysis results . . . 44

3.2.3 Historical analysis discussion . . . 47

3.3 Conclusions: New insight and policy recommendations . . . 49

4 Participatory research 51 4.1 A greenhouse grower survey . . . 52

4.1.1 Introduction, theoretical framework and methodology . . . 52

4.1.2 Survey results . . . 56

4.1.3 Survey discussion . . . 58

4.2 The evaluation tool and workshop . . . 60

4.2.1 Introduction, theoretical framework and methodology . . . 61

4.2.2 Workshop results. . . 65

4.2.3 Workshop discussion . . . 68

(9)

4.3 Conclusions: New insight and policy recommendations . . . 69

5 Agent-based modelling 71 5.1 An agent-based model of technology diffusion in the Westland-Oostland Greenport. . . 72

5.1.1 Introduction, theoretical framework and methodology . . . 72

5.2 The ABM experiment . . . 76

5.2.1 Model description . . . 76

5.2.2 Parameters. . . 86

5.2.3 Results. . . 87

5.2.4 Discussion. . . 95

5.3 Conclusions: New insight and policy recommendations . . . 98

6 Part I: Discussion and Conclusions 101 6.1 Summary of Part I insights . . . 101

6.2 Summary of Part I policy advice. . . 103

6.3 Conclusions from Part I. . . 105

6.4 Next steps following Part I. . . 106

II The TM questions: Applying TM to itself 109 7 Setting boundaries 111 7.1 The assumption. . . 112

7.2 The ABM experiment . . . 116

7.2.1 Hypotheses . . . 117 7.2.2 Parameters. . . 117 7.2.3 Results. . . 117 7.2.4 Discussion. . . 122 7.3 Conclusions. . . 124 8 Looking ahead 127 8.1 The assumption. . . 128

8.2 The ABM experiment . . . 133

8.2.1 Hypotheses . . . 137 8.2.2 Parameters. . . 137 8.2.3 Results. . . 138 8.2.4 Discussion. . . 142 8.3 Conclusions. . . 146 9 Experimenting 149 9.1 The assumption. . . 150

9.2 The ABM experiment . . . 156

9.2.1 Hypotheses . . . 160

9.2.2 Parameters. . . 161

9.2.3 Results. . . 161

9.2.4 Discussion. . . 167

(10)

10Learning 173

10.1The assumption. . . 174

10.2The ABM experiment . . . 179

10.2.1 Hypotheses . . . 184

10.2.2 Parameters. . . 184

10.2.3 Results. . . 185

10.2.4 Discussion. . . 190

10.3Conclusions. . . 192

11Part II: Discussion and Conclusions 195 11.1Revisiting the outcomes of Part I . . . 196

11.2Summary of Part II insights. . . 198

11.3Summary of Part II policy advice . . . 201

11.4Reflection. . . 202

11.5Future work. . . 203

References 205

A Summary 223

(11)
(12)

2.1 An ambiguous image of either a rabbit or a gull. . . 11

2.2 The demographic transition. . . 16

2.4 The MLP . . . 17

2.5 TM typologies:external influences and transition trajectories . . . 18

2.3 Carrying capacity S-curve . . . 18

2.6 The three pillars of sustainability . . . 22

3.1 Case study theoretical framework . . . 35

3.2 Total electrical CHP capacity in the greenport over time . . . 38

5.1 A sample network of greenhouse grower agents. . . 77

5.2 Initial distribution of heater technologies. . . 80

5.3 Agent inputs and outputs . . . 82

5.4 Control experiment: Mean possession of heaters . . . 88

5.5 Control experiment: Mean possession of irrigation technologies . . . 88

5.6 Control experiment: Mean possession of lighting technologies . . . 89

5.7 Control experiment: Standard deviation of heaters . . . 90

5.8 Control experiment: Standard deviation of irrigation technologies . . . 90

5.9 Control experiment: Standard deviation of lighting technologies . . . 91

5.10 Final credit balance in relation to company surface area . . . 91

5.11 Innovator probability & stubbornness: Mean H7 possession. . . 92

5.12 Opinion change rate & degree of neighbours: Early mean H7 possession . 93 5.13 Opinion change rate & degree of neighbours: Late mean H7 possession. . 93

7.1 Summary of mean possession over short time frame . . . 119

7.2 Summary of standard deviation over short time frame . . . 119

7.3 Summary of mean possession over long time frame. . . 121

7.4 Summary of standard deviation over long time frame. . . 121

8.1 Initial distribution of vehicle types . . . 134

8.2 Vehicle purchasing schema . . . 136

8.3 Uncompressed data: average electric and petrol vehicles over all cases . . 139

8.4 Uncompressed data: average electric vehicles by case . . . 140

8.5 Uncompressed data: average petrol and diesel vehicles by case. . . 141

8.6 ZIP file sizes: electric vehicles by lifespan . . . 142

8.7 ZIP file sizes: non-electric vehicles by type and lifespan . . . 143

8.8 ZIP file sizes: average sum of all vehicle types over all cases . . . 144

9.1 A typical landscape set up for the Adder model experiments. . . 156

(13)

9.2 Control experiment: Peaks reached and cost reductions . . . 163

9.3 Control experiment: An organic looking agent network . . . 163

9.4 Control experiment: Active repertoire size . . . 164

9.5 Control experiment: Offspring and uses per technology . . . 165

(14)

2.1 Three possible responses to system crisis . . . 23

2.2 Three modern governance theories . . . 25

2.3 Comparison of scientific method and TM cycle steps. . . 26

3.1 Key CHP diffusion factors . . . 41

3.2 SWOT analysis of sustainability in the Westland-Oostland Greenport . . . 47

3.3 Key greenhouse horticulture evolution factors . . . 47

4.1 Survey results: Adopter characteristics . . . 56

4.2 Survey results: Business motivations and company characteristics . . . . 56

4.3 Survey results: Economic and technological characteristics . . . 57

4.4 Survey results: Information sources . . . 57

4.5 Survey results: Calculating optimal decisions . . . 58

4.6 Evaluation tool and guiding criteria questions . . . 64

4.7 Evaluating the Westland-Oostland Greenport marketing event . . . 65

4.8 Evaluating the Green Campus Innovation Desk . . . 66

4.9 Using the portfolio worksheet on the results of the evaluation tool . . . 67

5.1 Greenhouse grower agent properties . . . 79

5.2 Technology lifespans and purchase prices . . . 80

5.3 Technology rankings by crop type . . . 81

5.4 Narrative summary . . . 82

5.5 Parameter settings for the control experiment . . . 86

5.6 Parameter settings for the non-control experiments . . . 87

7.1 Parameter settings for 250 and 2500 time step experiments . . . 118

8.1 Kolmogorov complexity of strings . . . 132

8.2 Parameter settings for experiment . . . 138

9.1 Example technology agent structure, product, cost and fitness . . . 158

9.2 Parameter settings for the control experiment . . . 161

9.3 Parameter settings for all non-control experiments . . . 162

9.4 Control experiment: All metrics and sub-metrics by case . . . 162

9.5 Non-control experiments: Problem solving success metrics . . . 166

9.6 Non-control experiments: Technology creation metrics . . . 166

9.7 Non-control experiments: Radical to incremental ratio metrics. . . 167

10.1 Parameter settings for experiment . . . 185

(15)

10.2 Least complex problems: problem solving efficiency and completeness. . 187 10.3 Most complex problems: problem solving efficiency and completeness . . 188 10.4 Intermediate complex problems: problem solving efficiency and

(16)

A PhD is awarded to those who can demonstrate that they can ‘do science’ to the satis-faction of those who have previously demonstrated such scientific abilities. Wooed by tales of academic glamour, scientific breakthroughs, intellectual excellence, or by a long career working indoors and with no heavy lifting, many wide-eyed hopefuls set out to prove that they too can do science. They may not begin with a clear idea of what it is to do science or how one demonstrates that science has been done, but they soon begin to learn. The successful will carry away more than a diploma; they will also bear the weight of naivité-busting experience and the scars of the scientific brotherhood’s initiation rites.

Like so many others that have gone before me, I stand at the end of the process, astonished at how different things look. I cannot emphasise enough the role of my pro-motors and supervisors in my transformation. Margot, Gerard and Igor, you have taken on the onerous and not infrequently painful duty of shepherding me through a long and arduous process. You have given your time, energy and attention to me as you read in-numerable drafts, shared feedback, and discussed painstaking details. I most certainly would not be where I am today without your efforts. I must also acknowledge the meta-morphosing influence of my committee; your time, effort and opinion have had an in-estimable effect on me and my work.

One important lesson of how to do science is that science cannot be done alone. Thus, I would specifically like to thank those of my colleagues who have been co-authors and editors: Reinier van der Veen, Emile Chappin, Janne M. Korhonen, Alfredas Chmieli-auskas, Zofia Lukszo, Koen van Dam, and (of course) Igor Nikolic and Gerard Dijkema. I would also like to thank the various students that I have supervised as they undertook tasks for my research: Wouter van den Berg, Luke Bergwerff, Thomas Mulder, Bas van IJzendoorn, Felipe Avancine, Theodoros Galanos, Aaron Hoffman, Noortje Schrauwen, Wei-Han Wu, Sam Wicki, Cornelis Eikelboom, Bert van Meeuwen, Jeroen van der Beek and Mark Vavier. Not all of that work led to publication, and not all of it contributed directly to this thesis, but I am grateful for the way that student effort has advanced my research and advanced by own abilities to teach and manage.

Less formally, there are many colleagues and friends whose impact on me, both in-side and outin-side of the work-osphere, has been no less important for its lack of formal-ity. The many Friday drinks, section dinners, World Cup betting pools, coffee machine chats, lunchtime discussions and random questions all proved enjoyable, challenging and enlightening and I will miss them. Gerben, in particular, has been as good friend and office-mate as one could hope for (Do I get a sticker for that?).

I also specifically want to thank my family. This includes family on both sides of the Atlantic and it also includes the many long-term, family-esque friends that I have

(17)

accumulated over my life. Your support and your faith in me has been invaluable, as were the opportunities you offered me to have some fun and to temporarily pull myself out of the thesis-pit into which I had not noticed that I had fallen. I don’t know how to express the depth of my gratitude, but I can offer you occasional baked goods.

Not all of my supporters are here to see this final step in my proving-that-I-can-science process, and so I raise a metaphorical glass to those who cannot join me in person. I especially want to thank all the women throughout my life who have been so strong and simultaneously so kind. You mean more to me that I could possibly say. Thank you.

Thank you. And thank you.

And, finally, I turn my attention to Jamie. I could spend a geologic age listing the ways that you have helped me, from cups of tea and gifts of chocolate to a good talking to or a silent hug when I needed them. But instead, I wrote you a haiku (because it is shorter).

Andamos juntos. Sin ti, no sería yo. Merci, mi amor.

(18)

1

I

NTRODUCTION

The Westland-Oostland Greenport is the largest concentrated cluster of interacting hor-ticultural industries in the Netherlands, where greenhouse horticulture forms an impor-tant part of the food supply, the economy (Breukers et al.,2008) and even the national culture (van der Veen,2012;van der Veen and Kasmire,2015). Like many others, the sec-tor now faces serious challenges from increased international competition, high labour prices, a reliance on insecure supplies of imported fossil fuels, and pressure to reduce its disproportionately high greenhouse gas emissions (European Environmental Agency, 2012). Facing any one of these challenges would be difficult, but facing them all at once makes everything much more complicated, especially when a possible solution to one challenge seems to make another worse. Investing in new machinery, for example, might make local producers more competitive and reduce labour costs but could increase fossil fuel reliance or greenhouse gas emissions.

All of these challenges, including the social and economic ones, can be understood as sustainability issues. Greenports appear to be complex adaptive systems (CAS) ( Chan-dler,2005;Fleming and Sorenson,2001;Holland,1996;Kasmire et al.,2011;Kauffman and Johnsen,1991;Kelly,2010;Newman,2003) with bottom-up behaviour that pervades the system but which cannot be reduced to component parts. Sustainability can be de-fined as a property that emerges within a CAS (Nikolic,2009) from the balance of eco-nomic, ecological and social values (Brundtland et al.,1987). In this light, the challenges facing the greenport appear to stem from a lack of balance, with some actors prioritising the preservation or expansion of some values at the expense of the others. A sustain-able greenport would prioritise an emergent and system-wide harmony over all three factors so that the currently valuable social and economic contributions remain or grow strong while environmental contributions improve to match. Furthermore, a sustain-able balance must also endure beyond the current threats by adapting to a new balance as needed in the future (Grin,2010;Meadowcroft,2000;Meadows et al.,2004). Green-ports are renowned for being innovative (Breukers et al.,2008), technologically advanced (Heichel,1976;Tomczak,2005;Walsingham et al.,1976), socially cohesive (van der Veen, 2012;van der Veen and Kasmire,2015) and open to sustainability (Verbong and Geels,

(19)

1

2007sary enduring balance.), all of which bodes well for their chances of achieving and maintaining the

neces-Theoretically, a CAS such as the Dutch greenhouse horticulture industry could spon-taneously transition from its current, unsustainable state of imbalance to a sustainable state of balance. Unfortunately, the need for sustainability is urgent (Grin,2010; Mead-ows et al.,2004) while the signs of an imminent transition to sustainability are weak (Lund and Freeston,2001;Rybach and Sanner,2000;TNO,2010). At the same time, emergent properties and CAS cannot be designed or governed top-down (Waldorp,1992), which puts the Westland-Oostland Greenport in a difficult situation; they cannot wait for a bottom-up transition to sustainability, but neither can they force that transition top-down. The Westland-Oostland Greenport thus asks ‘How can we manage a transition to sustainability?’ To answer that question, the greenport formed the Programma Du-urzame Greenport Westland-Oostland (DGWO)1with support from the European Re-gional Development Fund and the European Commission. Taking its cue from the way the question was framed, the DGWO embraces Transition Management (TM). TM is an approach to scientific research and policy formation that is popular in the Netherlands and beyond. TM aims to help socio-technical CAS steer themselves toward sustainabil-ity through a blend of bottom-up and top-down measures. As a research tool, TM is a four step cycle that is very much like the scientific method but that focusses on un-derstanding a CAS, identifying its sustainability problems and searching for signs of an emerging sustainable balance. As a policy tool, TM is a framework for reflexively set-ting and achieving goals in order to influence the CAS toward sustainability. These two foci complement each other so that a greater understanding raises opportunities for in-fluence and greater inin-fluence enhances further understanding. Importantly, TM is an iterative process so that the results of each cycle can be fed into the next, allowing in-dividual cycles to produce more understanding than influence, or vice versa, while also contributing to future understanding and influence.

TM certainly looks promising, with its appreciation of CAS, emergence, sustainabil-ity as a balance of pressures, and the need to avoid extremely top-down or bottom-up approaches. Yet, TM is a young and developing theory for which it is still not clear “what is the scope or feasibility of managing transitions” (Chiong Meza and Dijkema,2009). This means that in addition to (or in conjunction with) managing a transition to sustain-ability, TM must establish that transitions can be managed. TM attributes its remarkable popularity and progress so far (Chappin,2011) to reflexive applications of an iterative learning-by-doing four step cycle for CAS. By seeing itself as an object of study and ex-perimentation, TM hopes to improve understanding and influence for itself and for CAS with sustainability problems. By freely acknowledging its own imperfections, TM ex-pects every application to raise new questions, produce new insights and motivate new adjustments or modifications of its own theory and practice.

As this work was undertaken on behalf of the DGWO, it uses TM as an overarch-ing framework for exploroverarch-ing how a transition to sustainability might be managed in the Westland-Oostland Greenport. At the same time, this work also takes a second, higher level, reflexive approach that looks at the scope or feasibility of managing transitions, both within the greenport and elsewhere.

(20)

1

1.1.

R

ESEARCH

B

Yasking “How can we manage a transition to sustainability?”, the DGWO explic-itly embraces TM, its iterative searches for greater understanding and its blend of bottom-up and top-down influences. Thus, this main question can be broken down into a direct approach and a reflexive approach, each with distinct research sub-questions and appropriate research methods.

Research questions First, the main question and the contribution of TM can be taken at face value and applied directly to the DGWO question, producing two greenport spe-cific sub-question:

What new understanding and insight can be gained from applying TM to the Westland-Oostland Greenport?

What influence, policy recommendations and practical advice can be derived from the new understanding and insights gained by applying TM approach to the Westland-Oostland Greenport?

Answers to these sub-questions will necessarily be specific to the Westland-Oostland Greenport and any insights, recommendations and outcomes must be understood in that context. The answers may be useful, at least in an abstract way, outwith the green-port and the results could certainly shape future TM studies in other contexts. However, care should be taken to appreciate the specific context in which the results were pro-duced without extrapolating too much from this specific case to any other.

Second, TM’s emphasis on self-critical reflection and adaptive improvement allows the scope and feasibility of TM to be examined critically in two TM sub-questions:

What potentially problematic assumptions can be seen within the TM approach as it was applied to the Westland-Oostland Greenport?

What insights, further questions, or improvements for TM come from exploring these problematic assumptions?

Answers to these sub-questions apply to TM as a tool or process, and so are not contin-gent on the context of any specific TM cycle but are an inherent part of the critical self-examination that drives improvement within TM. However, improvements in TM may appear to undermine the results of earlier applications of TM. Why not do the critical self-examination before applying TM to the Westland-Oostland Greenport? The answer lies in the way that TM is fundamentally iterative and operates through ‘learning by do-ing’. The potential for improvements to TM can only be spotted in practice, with every application expected to reveal new possibilities for development that can be included in future applications. TM also emphasises that the problems of unsustainability are too urgent to wait for perfection before use; perfect understanding may be unattainable, but much can be achieved by taking advantage of mistakes, as well as successes, in the immediate term.

Research approach Both approaches to the main research question lends themselves to different methods and tools. First, the direct application of TM to the

(21)

Westland-1

Oostalnd Greenport is achieved through the use of various analysis techniques includ-ing a case study, a collaborative workshop, and an agent-based model. These TM tools

and methods are employed in conventional ways aimed at extracting new understanding about the Westland-Oostland Greenport, its current state, and its potential for sustain-ability. The insights thus gained then form the foundation for policy recommendations and practical advice that greenport can use to influence its path toward an enduring and balanced sustainable state.

Next, the reflexive application of TM to itself is achieved through a series of agent-based models that only abstractly represent the Westland-Oostland Greenport or similar CAS. Each model begins by identifying TM ideas or assumptions that can be understood as contradictory. Those conflicting ideas and assumptions are then used to define an agent-based model and one or more experiments within that model. The experimental results then reveal what, if anything, could change within TM in order to resolve the conflict, improve understanding, and increase the potential for influencing a transition to sustainability. In this way, the reflexive application of TM sets out to learn about the scope and feasibility of TM as well as ways that it might be improved.

1.2.

A

NTICIPATED RELEVANCE

,

SCIENTIFIC CONTRIBUTION AND

AUDIENCES

E

VERYscientific work comes with some expectations about how it could be useful, what might be revealed, and who is most likely to find it beneficial or interesting.

Relevance Work that addresses the issue of sustainability within the Westland-Oostland Greenport is obviously relevant to the greenport itself, but is also important for the Nether-lands and Europe. Even work that only shows some approach to be ineffective is impor-tant and relevant by preventing future wasted efforts and unnecessary delays. Thus, this thesis expects to contribute to the issue of sustainability within the Dutch greenhouse horticulture sector but hopes that the contribution may extend to other sectors in other countries too.

At the same time, the explicit focus on iteration and reflexivity embedded within TM offers another potential relevance; self-analysis is considered key to TM’s meteoric growth and popularity to date and so may be key to its future successes as well. By ex-plicitly using this reflexiveness to examine TM itself, this thesis expects to contribute directly to TM and indirectly to similar research and policy methods as well as wider sustainability science efforts.

Scientific Contribution This thesis anticipates two distinct but interrelated scientific contributions. The first contribution is to sustainability within the Westland-Oostland Greenport through the elucidation of new insights, behaviours, mechanisms, and re-lationships and their role in improving or impeding an emergent balance in that CAS. There may also be a scientific contribution to sustainability outside of the Dutch green-house horticulture sector, although that depends on how successfully the greenport-specific insights and policies can be abstracted or interpreted to apply more widely. The second scientific contribution is to TM, which will not escape unscathed from the

(22)

criti-1

cal examination of some of its internal problems and conflicts. TM makes no claims of perfection and instead embraces co-evolution between theory and practice, suggesting that it is well placed to reform itself in response to the criticism levelled against it here.

The intended audience consists of two main groups, although people outside of those groups may also benefit. First, scientists using TM are expected to be directly affected and so are part of the target audience. Scientists working on sustainability through other theories, frameworks, and methods may also find the ideas and approaches useful, al-though they are advised to be critical about how the insights might translate into their own research fields.

Second, policy-makers in the Dutch greenhouse horticulture sector, including lo-cal governments, business leaders, elected officials, consultants, advisers and experts, represent a part of the target audience. These policy-makers are in the unenviable po-sition of being asked to preserve what makes greenports valuable now while adding fur-ther value despite both current and future challenges. They are asked to work these mir-acles with limited finances, support and power while also being judged by the public and history. This thesis offers some practical advice on how sustainability, innovation or educational policies might be adapted in the short term and also offers some general advice on ideas to bear in mind over the longer term. Policy-makers working in other industries or other countries may find some valuable insights here, although they are encouraged to be cautious about applying the results of this work out of context.

1.3.

R

EADER

S

G

UIDE TO THIS THESIS

T

HEfollowing guide outlines what each part and chapter contributes and how these contributions relate to each other.

Chapter1Introduction lays out the problem to address, the approaches taken,

theories, examples and methodological tools used, and the research objectives, re-search questions, expected relevance, anticipated scientific contribution, intended audiences, as well as a reader’s guide.

PartI: Applying TM to the Westland-Oostland Greenport

Chapter2Understanding Transition Management presents relevant

back-ground knowledge on CAS, transitions, sustainability, modern governance, and TM itself, including the TM cycle.

Chapter3Historical analysis and case study introduces a case study of a

recent and rapid technology diffusion in the Westland-Oostland Greenport and a historical analysis of greenhouse horticulture evolution, as well as the new insights and policy recommendations that can be drawn from them. – Chapter4Participatory research presents a survey of greenhouse growers

and a goal setting and policy evaluation tool as well as the participatory work-shop in which that tool is used, before concluding with the new insights and practical policy advice motivated by those insights.

(23)

1

Chapterogy diffusion in the greenport and a set of experiments that explore the mod-5Agent-based modelling details an agent-based model of

technol-elled behaviour before wrapping up with the insights gained from the exper-iments and the policy recommendations that can be derived from them. – Chapter6Part I Discussions and Conclusions reflects on the first approach

to the main research question, summarises the insights gained from it, and recaps the practical advice derived from those insights. The chapter then dis-cusses the conclusions from this approach in relation to previous TM works and to possible future work addressing TM, sustainability and the greenport. The chapter wraps up Part I by briefly introducing how these conclusions re-late to the second approach to the main research question as addressed in Part II.

PartII: Applying TM to itself

Chapter7Setting boundaries introduces a TM assumption relating to

sys-tem boundaries, emergence and objectivity that might have sustainability consequences. The chapter then describes an agent-based model and set of experiments that explore the assumption before concluding with new in-sights and some ideas on how they might relate to policy-making.

Chapter8Looking ahead describes a TM assumption about the relationship

between transitions and stability and why that relationship matters for man-aging transitions. The chapter then details an agent-based model and exper-iments that investigate the assumption before summarising the new insights gained and how they might impact on policies and decisions.

Chapter9Experimenting presents a TM assumption regarding transition

ex-periments and describes how that assumption underpins TM efforts toward emergent sustainability. The chapter goes on to describe an agent-based model and series of experiments that examine the assumption before sum-marising the new insights obtained and their potential impact on policies designed to manage a transition to sustainability.

Chapter10Learning introduces a TM assumption about the role of learning

for complex problem solving and how that assumption influences the way TM works. The chapter then presents an agent-based model and set of ex-periments to explore the assumption before winding up with a summary of new insights and the potential policy repercussions.

Chapter11Part II Discussion and Conclusions begins by reminding the reader of

the first approach to the main research questions and the conclusions gained by applying TM to the greenport in Part I before reviewing the second approach to the main research question taken throughout Part II. The insights gained through applying TM to itself are summarised and then contextualised to produce some practical advice and policy recommendations. The chapter and part concludes with a reflection on the work as a whole, a discussion of implications for interpret-ing and applyinterpret-ing the work, and some ideas for future research that may advance on the themes and ideas presented.

(24)

I

T

HE GREENPORT QUESTIONS

:

A

PPLYING

TM

TO THE

W

ESTL AND

-O

OSTL AND

G

REENPORT

(25)
(26)

2

A

PPLYING

TM

TO THE

W

ESTL AND

-O

OSTL AND

G

REENPORT

: U

NDERSTANDING

T

RANSITION

M

ANAGEMENT

Abandon the urge to simplify everything, to look for formulas and easy answers, and to begin to think multidimensionally. . . to appreciate the fact that life is complex. M. Scott Peck

All of the important underlying concepts that led to the development of Transition Man-agement (TM) must be introduced and explained before it can be applied to the Westland-Oostland Greenport. This chapter therefore contains a literature review that explores the ideas behind CAS, transitions, sustainability, and modern governance, tracks the devel-opment and interaction of these various theories and ideas, and defines the relevant vo-cabulary. The chapter then concludes with an introduction to TM itself and the TM cycle, including all of the key concepts, goals, and ideas that underpin how TM is used.

Parts of this chapter have been published previously (Kasmire et al.,2012a,b;Nikolic and Kasmire,2012;Nikolic et al.,2012)

(27)

2

2.1.

I

MPORTANT UNDERLYING CONCEPTS

T

RANSITIONMANAGEMENT(TM) was created to help stimulate, steer and shape tran-sitions to sustainability within socio-technical CAS, which means that TM depends on the concepts of transitions, sustainability, and CAS, as well as a few other concepts that cross between these categories. These important concepts are best explained by starting with CAS.

2.1.1.

CAS

Historically, systems were understood to be closed, simple and reducible, meaning that all the relevant parts and interactions of any system could be completely separated from everything else, that those parts and interactions were not meaningfully divisible, and that understanding a system was a matter of understanding all of its component parts. A full and complete system definition was understood to be a description of the sys-tem’s exact initial conditions and all applicable laws, which was considered sufficient to predict the entire system’s history. Such systems were also thought to be control-lable through their initial conditions, which is to say that manipulating the initial con-ditions was equivalent to managing the system’s future. Because the entire system his-tory (or future, depending on point of view) was considered contained within the sys-tem description, and because such syssys-tems were thought to inevitably settle into a time-independent equilibrium, systems were typically considered reversible. However, TM focusses on the more recently identified complex adaptive systems (CAS) which are un-like the closed, simple, reversible and reducible systems classically studied by science and which are instead full of unpredictable, hard-to-manage and impossible to fully de-fine social and technological components (as well as economic, psychological, physical, cultural, and just about any other adjective that can be applied to systems with people in).

CAS still use system definitions, but those definitions are now considered to be ap-proximate models of the system rather than exhaustive descriptions. Such system defi-nitions, like all models, are at least two steps removed from reality1(Nikolic and Kasmire, 2012) which makes them very difficult to describe. CAS definitions are also expected to display several features that make them different from the closed, simple definitions of previous system study; CAS are observer dependent idealisations, have multiple, inter-dependent and organised components, boundaries, an environment, states, and display non-trivial behaviour (Ryan,2008).

Observer dependent idealisations are a consequence of the way that one CAS can be seen in different and possible mutually exclusive ways by different observers (See Figure 2.1). Every observer is somewhere, meaning that no observer can every observe from a theoretically objective and observer-independent nowhere (Haraway,1988). Systems are also nested, meaning that they are composed of smaller sub-systems and embedded in larger systems. This forces observers, who cannot chose the somewhere from which they observe, to choose what system to observe, what scale at which to observe it, and which

1Every perception or interpretation of a system is a mental model, removed from reality by one step.

For-malised descriptions, models and representations that can be shared with others are models of models, re-moved from reality by a further step.

(28)

2

patterns or behaviour are interesting enough to become the centre of observant atten-tion. For example, within what is ostensibly the same system, one observer might see individuals interacting economically over the course of months or years while another sees corporations or countries interacting globally over decades. Observers also have their own unique (and not always chosen) perspective, which allows even very simple systems, like a single chair, to be perceived in different ways. For example, one observer might see that chair as an example of post modernism while another sees a variable lead-ing to improved classroom behaviour, and yet another sees an object defined by prop-erties such as mass, strength and volume. No single observer captures every possible detail about a system and instead ignores those that are irrelevant to the chosen scale, scope or perspective. Ignoring or attending to the different details of a system creates idealisations or abstractions based on utility, which means that one observer dependent idealisation can be as valid as another, albeit possibly for a different purpose.

Figure 2.1: Observer dependence: the figure can be seen as a rabbit or a gull, or both in alternation, but not both at the same time.

The multiple, interdependent, and organised components of a CAS interact in many ways. Matter, energy and information flow from some system components to others, possibly being changed in the process. For example, the var-ious organisms in a food web prey upon, mate with, or avoid each other which can pass nutri-ents and genetic material between system compo-nents. Interactions between system components can be straightforward or not, accidental or pur-poseful, and desirable or not, but are organised when system component interactions are not uni-formly frequent or bi-directional. The organisms in an ecosystem are not all equally likely to eat and be eaten by each other, nor can they reproduce with all other organisms.

System boundaries used to be considered the impenetrable demarcations of timeless, atomic components, but there is no known system whose outer system boundary is not penetrated by energy, matter or information that influences its internal workings, whose components are not divisible into further sub-divisions, and whose interactions are completely independent of time. Despite this, seeing system descriptions as useful abstractions or models means that they can have system boundaries even when the sys-tems they describe do not. These boundaries clarify which system components, interac-tions and influencing factors are considered important enough to observe and, if done well, are drawn in full awareness that they represent arbitrary groupings of apparently useful subsets of a larger system (or systems).

The in-out boundary of a CAS is porous so that recognised types of energy, matter and

information can cross in and out, usually with expectations of the volume and direction of each recognised cross-boundary flow. Like the boundary itself, the set of recognised boundary-crossers should not be considered absolute or definite but as an observer dependent idealisation of the anticipated or observed flows.

(29)

2

The focal level boundary captures the stable organisations that arise at every system

level as a result of frequent interactions between similar components at that level. By frequently interacting, each recognisable cluster or unit then interacts with similar clusters or units at a higher (and slower) level (Holling,2001). Because the interactions at each level share a time scale (or are recognised by their shared time scale), interactions at other levels seem remote and simplified, so “three levels of a hierarchy, one up and one down, are generally considered sufficient for analysis of the focal level” (Ryan,2008). The components at every level of analysis are made up of smaller and faster interacting components at a lower level, but setting the lower bound of the system description effectively renders the components at that level as atomic.

The time boundary captures the way that CAS interactions reveal recognisable clusters

or units in time as well as in space or system level. The time boundary defines the shortest and longest behaviours to be observed within the system, bearing in mind that long time frames lose or blur fast moving behaviours at lower lev-els while short time frames cannot describe long or slow moving interactions and relationships. Like the focal-level system boundary, the time boundary simpli-fies and atomises some components, namely those present at the beginning of the time frame, by obscuring their composite nature and development.

Complicating the matter, systems can be nested in time, physical space and social relations, among other possible ways. Every system, and every component of every sys-tem, can belong to more than one recognisable cluster, unit or larger system and can be decomposed into more than one set of parts or sub-systems, which means that there is no right way to set the system boundaries2.

The environment of a CAS is all the ‘rest’ of the world outside the observer dependent system boundaries, including what is outwith the focal level and time boundaries. Al-though the environment contains everything that is not specifically included within the system boundaries, it is usually abstracted into a set of variables, parameters or sources of influence that matter to the system, such as the permitted boundary-crossers. Thus, the price of natural gas, weather conditions, and the popularity of a given celebrity on so-cial media are all parts of the environment for a modern socio-technical CAS, but would not all be equally likely to feature in the defined environment of a CAS that TM might be interested in.

The state of a CAS comprises the enduring properties or patterns that attract an ob-server’s attention and make the system worth observing and defining. The system bound-aries define all possible system properties, which are all the ways that a system could possibly be at any given moment3. Every unique combination of all possible properties can be understood as a single point in a high dimensional space known as a system state

2And therefore no wrong way to set system boundaries. 3The time boundary also defines what counts as a moment.

(30)

2

space. The system’s current state, or how it is, represents a single point within that sys-tem state space while the entire set of points within that syssys-tem state space represents the entire set of ways that the system is or could be.

From one moment to the next, system properties can remain unchanged, change in regular, repetitive or structured ways or change in apparently unpredictable ways. These changes (and absences of change) mean that the single point representing the current system state moves around within the system state space. The path of the point repre-senting the current system state over time may create recognisable and enduring pat-terns, although these too are observer dependent. The entire set of ‘current moment states’ that contribute to an enduring pattern can be understood as a stable State (note the capital letter).

Stability can be broken down into three important and different concepts: constancy, robustness and resilience(Hansson and Helgesson,2003). Constancy is when a system State remains (largely) unchanged and always refers to an observed period of that sys-tem’s actual history(Hansson and Helgesson,2003). Robustness is when a system State remains (largely) unchanged in the face of disturbances, but can refer to observations of the system’s history or to a hypothetical period in the future (Hansson and Helgesson, 2003). Homoeostasis is a classic example of robustness wherein living systems maintain a relatively constant body temperature across a wide range of environmental conditions while thermostats, cruise controls or safety regulators are examples of technological ro-bustness. Resilience is when a system returns to its (approximate) original State after a disturbance (Hansson and Helgesson,2003) and, like robustness, can refer to observa-tions of past system behaviour or to expected future behaviour. Constancy, robustness and resilience do not always go together and system States can be considered stable if they display only one of the three.

The non-trivial behaviour displayed by CAS is caused by the way its multiple, inter-dependent and organised components react to each other across system levels, to the environment, and to themselves over time. These recursive and reflexive interactions create behaviour that is very different from the trivial behaviour or invariant mappings between system inputs and outputs (Von Foerster,1972) expected in simple, closed, sin-gle level and reversible or time-independent systems. Although not entirely separable, there are several non-trivial behaviours that are especially important in CAS, which are introduced here but which will become clearer as they are used extensively in sections 2.1.2and2.1.3.

Emergence and emergent properties are the seemingly magical appearance of new

char-acteristics or phenomena at one system level logically resulting from interactions at lower levels (Crutchfield,1994;Morin,1999). Examples include traffic jams (Nagel and Paczuski,1995), bird murmurations (Cucker and Smale,2007), human consciousness (Dennett,1996), the decisions and policies of institutions, govern-ments or corporate boards (Cohen et al.,1972;Lindblom et al.,1980), sustainabil-ity (Nikolic,2009), pollution, house price inflation and many other externalities (Funtowicz et al.,1997). Because the high level emergent properties are so much easier to recognise and understand than the myriad lower level interactions that they come from, emergence is often treated as a ‘black box’.

(31)

2

Importantly, emergence is not reducible because it is ‘more than the sum of its parts’. Emergent structures or phenomena cannot be de-constructed into com-ponent parts (Jennings,2000), would not arise in isolation (Morin,1999), may be lost if the system is broken apart, and may be lost in the parts that are removed4. Counter-intuitively, emergent structures, properties or phenomena can still ap-pear, disapap-pear, or change with the gain or loss of a single component or interac-tion. Irreducible, bottom-up processes like emergence have no centralised control or single cause, but because emergent properties can be desirable or undesirable, many decisions (especially from authorities) are made in an attempt to manage or influence emergence.

Adaptation is improvement over time in relation to environmental aspects that can be

physical, economic, technological, or social, among many others. Adaptation is a domain-neutral algorithmic process (and thus not restricted to biology) that oc-curs whenever a system displays innovation (or variation), retention (or inheri-tance), and selection (or competition) (Blackmore,2000;Dawkins,1976;Dennett, 1996;Nikolic and Kasmire,2012) regardless of the specific means through which the innovation, retention and selection act (David,2000;Jablonka,2000). Innova-tion is the ability to create novelty or differences in some copying process while retention is the ability to create reliable copies. Although seemingly at odds, nei-ther innovation nor retention are absolute so that the copies are more similar to the original than they are to unrelated others without being identical. Selection interacts with innovation and retention to determine which innovations are provements and therefore disproportionately retained. All three are equally im-portant.

Adaptation entails innovating on the basis of what already exists in ways that are improvements (at least in the short term), which means adaptation cannot identify the best possible approach and is shaped by what counted as an improvement in the past. Adaptations from very different origins sometimes converge on a single ‘good solution’ because selection can drive diverse adaptations in similar ways. What counts as a good solution can change over time in response to changes in the environment or to changing adaptations themselves. Referring to adaptation as co-evolution emphasises the way that adaptations are never isolated from each other or from the environment and that competition can ratchet up what counts as a ‘good solution’ for all the competitors (Futuyma,1983;Jantzen,1980;Thompson, 1994).

Chaos includes feedback and feedforward loops and arises in CAS and other

determin-istic systems when non-linear rules are applied iteratively. When outputs are re-turned as inputs in this way, a system’s existing patterns, interactions, organisa-tions, structures and emergent properties can become amplified and can suddenly send the output in unexpected directions5. By producing new output via

non-4Consider how the emergent property of living is lost if an organism is dissected and how an amputated part

ceases to be alive.

5Biological examples of such chaotic amplification of structures or interactions include peacock tails, cheetah

(32)

2

linear rules and by returning the output as the next input, chaos entails innovation and retention. Adding selection makes the system adaptive, which means that all adaptive systems are chaotic but not all chaotic systems are adaptive.

Being non-linear, chaotic behaviours are very sensitive to initial conditions, irre-ducible and intractable, all of which makes them impossible to exactly or specifi-cally predict6. At the same time, the non-linearity and iteration of chaos produces characteristic structures, such as the attractors and repellors that create stable States, enduring adaptive ‘good solutions’, and the self-similarity (or scale invari-ance) underpinning system nestedness. These characteristic structures drive con-vergent adaptation, patterned behaviour and stability, allowing for some limited predictions relating to a system’s general or short term behaviour. Nevertheless, the system can suddenly shift from one stable behaviour, pattern or State to an-other in an attractor shift, leaving the whole system intractable and unpredictable over the long term.

Self-organisation describes the distributed, chaotic (and often adaptive) process by which

an emergent structure, pattern or State becomes stable, durable and self-reinforcing by building itself from within itself, without any central or external authority (Kay, 2002;Prigogine and Stengers,1984). For example, embryos self-assemble into fully functional organisms from a single fertilised cell (Campbell,2002), societies develop from limits on choice and behaviour that drive further limits on choice and behaviour (Luhmann,1995), and the formation of crystals, galaxies, micelles and cellular automata all promote their own further growth. Self-organisation can apply in time as well as space in what are known as self-organising critical-ities. Attempts to influence socio-technical systems may try to piggyback self-organisation so that the desired change ‘makes itself’ after only a small initial ef-fort. However, even very stable and enduring examples of self-organisation must adapt or perish if the rest of the system and its environment undergo sufficient change.

2.1.2.

T

RANSITIONS

Transitions are a shift from one dynamic equilibrium or stable State to another (Chappin, 2011;Chiong Meza,2012;Rotmans and Loorbach,2010) as measured by some selected indicator of change. At first, transitions were merely a descriptive label attached to ob-servations (Szreter,1993) of an S-shaped pattern of population growth (See Figure2.2). Later, the same basic S-shaped pattern was applied to the growth of many other things, like infections, ideas, technologies, processes or fashions, in what became known as dif-fusions. After many years of exhaustive research, the theories behind transitions and diffusions gained explanatory as well as descriptive power which offered hope that they may eventually be used for prediction or even for management and control (Davis,1945; Notestein,1945). In this context, transitions are best understood in terms of how they relate to the various non-trivial behaviours displayed in CAS.

6This unpredictability is very different than the unpredictability of randomness, which has no cause,

transfor-mational rule, or information content. Randomness is actually very hard to produce and the only suspected source is the decay of radioactive atoms driven by quantum fluctuations (Green,1981).

(33)

2

Figure 2.2: The demographic transition is a clear S-shaped curve in total population numbers. The dy-namic equilibria on either side of the curve are pro-duced when birth and death rates roughly match, but the transition is created when a drop in the birth rate lags behind a similar drop in in the death rate.

Transitions are emergent for two rea-sons: the dynamic equilibria on either side of the change are determined by high level, system-wide properties that emerge from the specifics of the system definition while the shift is a large-scale, high-level pattern or behaviour that emerges from lower level interacting components or be-haviours.

First, the dynamic equilibria are the relatively stable States on either side of the transition determined by the system’s carrying capacity for a chosen indica-tor of change. Carrying capacity can be roughly understood as the theoreti-cal maximal robust number of some sys-tem component given its relation to other components within a given system (See

Figure2.3). For populations this means the largest population that can be maintained within some region, calculated by dividing the regional productivity by per capita re-quirements (Dewar,1984). When there is an excess of all necessary resources the lation is considered to be below carrying capacity and will expand quickly. As the popu-lation expands, competition will increase for those resources that are most scarce until the population reaches the carrying capacity and stops growing. Time lags and complex system interactions mean that the population may sometimes exceed carrying capacity, and the intense competition for the scarce resources will produce crisis and population decline. Carrying capacity works much the same way for innovation adoptions, with market penetration figures in place of population numbers, markets in place of the ‘pro-ductive region’ and potential adopters as the ‘(scarce) resources’ that innovations com-pete to access. All calculations of carrying capacity require a system definition to deter-mine the region (or market), its resources and their availability, and the competitors and their resource requirements. Thus, carrying capacity and any dynamic equilibrium in re-lation to that carrying capacity will emerge from the system definition; different system boundaries, system components or environments produce different carrying capacities. Naturally, the (typically) S-shaped pattern of change between two emergent States is also emergent as “[e]very S-shaped curve is an aggregate of underlying curves, and the end-point of any transition curve may be the beginning of the next transition curve.” (Rotmans and Loorbach,2010, p. 129). Not only are the curves made up of smaller curves, but they are produced from lower level interactions between innovations (Geels and Schot,2010), innovation diffusions (Rogers,1995), business cycles (Schumpeter, 1934), internal adaptations (Omran,1971) and system boundary crossing external in-fluences (Caldwell,1976;Suarez and Oliva,2005), all of which are, in turn, made up of lower level components. Innovations (or “purposed systems” including technologies, processes, business models, and behaviours (Arthur,2009)) are composed of multiple parts in various possible arrangements (Arthur,2009;Henderson and Clark,1990).

(34)

Dif-2

fusions are high level phenomena made up of the low level adoption decisions of many individuals, beginning with the rare innovators who adopt first, followed by the rela-tively rare early adopters who adopt next and then by the more common early majority and late majority adopter before finally being followed by the relatively rare laggards who only adopt after the innovation is well established and understood (Rogers,1995). Similarly, business cycles are high level patterns composed of lower level, smaller, faster cycles, and internal adaptations are the emergent results of multiple co-evolving parts and pressures. Even external influences can be understood as having parts and struc-ture (see Figure2.5(a)) with varying frequency, amplitude, speed and scope (Geels and Schot,2007;Suarez and Oliva,2005). This emergence and system nestedness is captured in the multi-level perspective (MLP) (Geels,2002), which casts transitions as macro (or landscape) level changes arising from interactions within and between the environment and the meso (or regime) and micro (or niche) levels that are “so powerful that they re-sult in mal-adjustments, tensions, and lack of synchronicities . . . [that] create windows of opportunity for transitions.” (Geels and Schot,2010, p. 42). Emergence can arise, change or disappear with the gain, loss or alteration of a single component or interaction, which can theoretically begin at any system level.

Figure 2.4: The Multi-level Perspective, showing how the S-curve of transitions relates to system levels and evolutionary mechanisms (Geels,2002).

Transitions are adaptive because the shift ends in a more optimally ordered or struc-tured State (Rotmans and Loorbach,2010) that is better adapted to its environment (Nikolic et al.,2009). Evolving systems are always as well adapted as possible, mean-ing that the post-transition State is adaptively better, rather than just different, through a change in the system’s carrying capacity. Carrying capacity can change as a result of a new or altered feature of the environment, such as a sudden abundance or dearth of some resource that used to cross into the system in a consistent way. Small-scale or short lived environmental changes are unlikely to perturb the system enough to overcome the

(35)

2

(a) Typology of external influences (b) Transition trajectory typology

Figure 2.5: Some TM research has produced typologies describing how specific shocks, disruptive change and avalanche change are the external influences (2.5(a)) most likely to be involved in a transition (Geels and Schot,

2007) or how resources and coordination can vary (2.5(b)) to produce different transition trajectories (Berkhout et al.,2004).

stability and robustness of its carrying capacity, but large-scale, prolonged, multiple or mutually reinforcing changes might (See Figure2.5(a)). However, emergence within a system can also produce changes in carrying capacity.

Figure 2.3: The number of individuals at any given time rises in as S-shaped curve up to the carrying capacity, or the maximum number of individuals that the envi-ronment can support.

Innovations create societal moderni-sation (Schot, 2003), economic change and progress (Schumpeter, 1934) and adaptive transitions7(Caldwell,1976) by raising carrying capacity through in-creases to productivity, access to new resources, resource availability, resource use efficiency or changes to resource re-quirements. Innovations can enter a sys-tem via the in-out boundary (Caldwell, 1976) or by emerging through the focal level or time boundaries. An innovation’s value and contribution are high level fea-tures that emerge as a consequence of be-ing built recursively out of pre-existbe-ing in-novations (Arthur,2009;Henderson and Clark, 1990) and from interactions and

co-evolution with the system in which they are embedded (Arthur,2009;Kelly,2010; Kline et al.,1985;Kline and Model,1997).

Innovations are almost universally classified as either radical or incremental accord-ing to various evaluations of an innovation’s novelty, uniqueness, conceptual advance and impact on the status quo or on the future (Ahuja and Morris Lampert,2001;Dahlin and Behrens,2005;Fleming,2001;Henderson and Clark,1990;Poel,2003). Although ev-eryone has an intuitive idea of which innovations are ground breaking and radical and

7Transitions are sometimes referred to as ‘system innovations’ to emphasise the fact that they are very high

(36)

2

which are merely incremental modifications or revisions, the differences are often sur-prisingly context and observer dependent. Almost all studies use ex post facto measures of their impact (Dahlin and Behrens,2005) or estimates of their expected impact (Koberg et al.,2003;Schumpeter,1934;Stevens and Burley,2003) rather than (or in addition to) the messy and subjective measures of novelty, uniqueness or creativity. This tendency suggests that an innovation’s impact may be easier to detect and measure because those impacts are high level properties that emerge from the innovation’s interactions with the wider system while novelty, uniqueness and conceptual advances are low level (and therefore less detectable) properties of the innovation itself.

By raising carrying capacity, innovations create a resource excess that puts the sys-tem below the newly elevated carrying capacity. The selection pressures that previously checked population growth are suddenly weak and inconsistent in relation to the new carrying capacity, although they eventually grow strong as the population increases up to the new limit. Incremental innovations only raise the carrying capacity a small amount, meaning that selection is only slightly and temporarily weakened. Radical innovations produce large increases in carrying capacity that drastically weaken the selection pres-sures to create ‘windows of opportunity’ for transitions to emerge (Geels and Schot, 2010). Innovation is thus understood to be directly pitted against selection so that in-dividual innovations, especially radical ones, become ‘the seeds of transitions” (Geels and Schot,2010).

Focussing on the innovation in a system tends to stress creativity, ingenuity, and logic, while a focus on the selection downplays the intentional and gives the sense that everything is down to criteria in the selection environment, stochastic trajectories, blind luck, misunderstandings, copying errors, or random curiosity (Geels and Schot,2010). However, selection, which is definitely not random, represents “the environment into which these seeds are sown [which] is, of course, the main determinant of whether they will sprout” (Mokyr,1990) and how they must adapt after sprouting. Selection only looks stochastic or random because the environment is almost invisible in classic S-curve models (Geroski,2000) which only examine a simple selection change coming from a single innovation, according to a single indicator of system change and within a rela-tively closed and simple system. More recent models account for the complexity and co-evolution (Geroski,2000) with multiple system levels and historical, environmental, political, technological, psychological and biological contexts (Rogers,1995) as well as porous system boundaries. Some of these models produce typologies that better capture the complex and co-evolving interactions between selection and innovation in various idealised transition patterns (Berkhout et al.,2004;Geels and Kemp,2000,2007;Geels and Schot,2010). The most common type of transition, or at least the most commonly studied, are ‘emergent transformations’, with primarily bottom-up change and a large reliance on external resources (See Figure2.5(b)) (Chappin,2011), although real transi-tions can display features of more than one trajectory (Geels and Schot,2010).

Transitions are also chaotic because the iteration that allows for adaptation also al-lows for looped interactions. Innovations raise carrying capacity and enable population growth, but large and dense populations are more likely to create innovations (Boserup, 2005;Malthus,1966;Vasey,2002). Thus, population size and innovation are mutually

(37)

2

linked in looped interactions (Johnson,2010) known as reflexive downward causation, meaning that they are both the cause and effect of themselves and each other (Kim, 1999). When adaptations and the drivers behind those adaptations are linked in these mutually interacting loops, they can create various kinds of stability such as a Red Queen’s stalemate or evolutionary arms race (Carroll,2006), a co-evolutionary steady State ( Schaf-fer and Rosenzweig,1978), or robust dynamic equilibria (Taylor and Jonker,1978).

These stable States are not accidental or anomalous, but are the observable out-comes of one the many, potentially dynamic characteristic structures that emerge within CAS system state spaces. These characteristic structures are points or sets of points in the system state space that are created by selection and that attract or repel the sys-tem and are called, appropriately enough, attractors and repellers. Attractors are sur-rounded by a ‘basin of attraction’8that, once entered, are almost totally inescapable. Being emergent features of a CAS, the location, size and shape of characteristic struc-tures are unpredictable, but a system captured by an attractor is strongly affected by the selection pressures that created that attractor. The system responds to the attrac-tor/selection with patterned or stable behaviour that not only reveals the presence of the attractor/selection but also some of its features. A non-dynamic equilibrium indicates the very strong selection of a fixed-point attractor (Rotmans and Loorbach,2010) while dynamic equilibria and periodic behaviour indicate the strong but not irregular selection of multiple and/or variable point attractors. Aperiodic or chaotic behaviour that cannot be statistically described without ambiguity but that still seems stable or patterned, such as a climate (Lorenz,1964)9, suggests a strange attractor. A system’s movement within a (strange) attractor may still be intractable in specific terms, but general predictability is improved in that the system can be expected to remain trapped within the attractor, at least in the short term.

Despite this, chaotic systems can suddenly become unstable and move with an un-settling and unexpected inescapability in what is called an attractor change. For exam-ple, large crowds sometimes erupt into riots, crushes, or stampedes under conditions that are not obviously different from the conditions in which crowds of the same size act peaceably. Transitions are periods of instability in between one State or dynamic equi-librium and another, which means they are also attractor changes (Grin,2010;Rotmans and Loorbach,2010). Attractor changes are non-random and fully deterministic, but any iteration capable of producing chaotic behaviour will be sensitive to initial condi-tions. This sensitivity can magnify tiny differences to the point that they produce very unexpected outcomes in what is commonly known as the butterfly effect and is another way of saying that emergent phenomena can suddenly appear, disappear or change fol-lowing the inclusion, removal or alteration of a single innovation, external influences, low level emergence or other minute differences can be magnified to the point that they produce very unexpected outcomes. Thus, the stability of dynamic equilibria and other stable States is linked to strong selection pressures while the instability, unpredictabil-ity and sensitivunpredictabil-ity to small differences means that transitions are linked to innovation (Rotmans and Loorbach,2010). Furthermore, since selection and innovation are at odds with each other, States and dynamic equilibria are linked to periods of weak innovation

8Repellers may also have ‘basins of repulsion’, although these do not seem to be well studied. 9This surprisingly short paper can be interpreted to show that there is no such thing as a climate.

Cytaty

Powiązane dokumenty

The n × n matrix has a determinant which is the generalization of this rule of alternating sums determinant of such submatrices, mutiplied by the entry that is in the row and

— Apollos, the second brother of Nicanor, appears only once (No.. After the death, or retirement, of Nicanor in A.D. 50, Peteharpochrates took his place, being possibly assisted

Po dwudziestowiecznych rewolucjach w sprawach wojskowych – nuklearnej i następującej po niej rewolucji informacyjnej, w XXI wieku można wymienić nowe obszary, które w dającym

Keeping the type of option constant, in-the-money options experience the largest absolute change in value and out-of-the-money options the smallest absolute change in

² If the data values are grouped in classes on a frequency table or column graph, do we still know what the highest and lowest values are..

• “Nowy Sącz Experiment” in the end of 50’s and 60’s years and its influence on city's innovation,.. • 7 economical “tigers” – there is always somebody behind

Ex- plosive mixtures of dust and air may form during transport (e.g. in bucket elevators) and during the storage of raw mate- rials such as cereals, sugar and flour. An explosion

Gimnazjum z Polskim Językiem Nauczania w Czeskim Cieszynie jako znaczący ośrodek krzewienia kultury muzycznej na Zaolziu.. [...] artystyczne wychowanie, czy też lepiej wychowanie