Content uploaded by Lara Mia Herrmann
Author content
All content in this area was uploaded by Lara Mia Herrmann on Feb 24, 2017
Content may be subject to copyright.
The Non-knowledge Map for Decisions
A Heuristic for the Exploration and Creation of Non-Knowledge Literacy
I
do
not
know
what
else
to
put
on
the
cover.
Master thesis to gain the Master of Science degree (MSc) in
Global Change Management
Faculty of Forest and Environment at Eberswalde University of Sustainable Development (University
of Applied Sciences), Germany
Written by
Lara Mia Herrmann (BSc)
Born 9th December 1989
laraherrmann@web.de
Supervisors and Evaluators:
Prof Dr Pierre L. Ibisch, HNEE, Germany
Dr Peter R. Hobson, Writtle College, United Kingdom
Berlin, Germany
13th November 2015
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
2
This is a request for conceptual (non-)knowledge of higher order.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
3
I. Abstract
Decision making under uncertainty is frequently discussed in literature, but rarely considers other
forms of non-knowledge. Literature on non-knowledge does not provide practical classifications for
decision unknowns. This thesis seeks to find patterns in decision related unknowns. These patterns
shall serve to derive elements and principles of a competent handling of non-knowledge, which
would allow making non-knowledge literacy teachable. The post-normal scientific investigation at
hand is structured along the four phases of the panarchy (Gunderson and Holling, 2002). It uses an
econical approach to (non-)knowledge management (Hobson and Ibisch, 2012). This means that
principles of scientific exploration are enriched with principles of (non-)knowledge management
observed in ecosystems. A multidimensional non-knowledge map is used to systematically analyse
unknowns from 40 work decisions of Global Change Managers. A post analysis review, the fourth
phase of the panarchy, has reorganised the dimensions of non-knowledge in a simple heuristic that
can be written in code. This simple heuristic orders points in time and assigns them an intelligible
(non-)knowledge quality indicator. Initial results of patterns that emerge are presented. Handling
non-knowledge is an inherent capacity of humans; at work this handling often remains implicit. As a
cultural effort, non-knowledge literacy would be a competence that invites changes in mind-set,
behaviour and institutional structures. Some principles of non-knowledge literacy are collected for
further investigation. These could be explored by adding variables of handling and evaluation of
unknowns to the given heuristic. Non-knowledge literate professionals would be able to understand,
explain and strategically handle their unknowns.
Keywords: non-knowledge, decision making, heuristics, econics, post-normal science, panarchy
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
4
II. Zusammenfassung (German Abstract)
Titel: Die Nichtwissenskarte für Entscheidungen – eine Heuristik für die Erforschung und Bildung
von Nichtwissenskompetenz
Entscheidungsfindung unter Unsicherheiten wird in der Literatur häufig thematisiert, nimmt jedoch
selten Bezug auf andere Nichtwissensformen (z.B. Blinde Flecken oder Wissenslücken). Literatur zu
Nichtwissen liefert keine praktisch anwendbaren Kategorien für entscheidungsbezogene
Unbekannte. Diese Arbeit strebt an, Muster in entscheidungsbezogenen Unbekannten zu entdecken.
Diese Muster sollen dazu verwendet werden, Elemente und Prinzipien für einen kompetenten
Umgang mit Nichtwissen abzuleiten. Das würde eine „Alphabetisierung“ im Umgang mit Nichtwissen
ermöglichen. Die vorliegende post-normale wissenschaftliche Untersuchung wird entlang der vier
Phasen der Panarchy (Gunderson and Holling, 2002) strukturiert. Dieser ökonische Ansatz erweitert
bestehende Prinzipien der wissenschaftlichen Forschung mit Prinzipien der Wissensverwaltung
welche in Ökosystemen beobachtbar sind. Eine mehrdimensionale Nichtwissenskarte wird genutzt
um spezifische Ungewusste von 40 Arbeitsentscheidungen von Global Change Managern
systematisch zu untersuchen. Nach der Diskussion der vorherigen Analyse werden die Dimensionen
der Nichtwissenskarte in einer einfachen, kodierbaren, Heuristik neu zusammengestellt. Diese
einfache Heuristik ordnet entscheidungsrelevante Zeitpunkte und weist ihnen jeweils einen
verständlichen Indikator der (Nicht-) Wissensqualität zu. Erste Ergebnisse der erscheinenden Muster
werden präsentiert. Umgang mit Nichtwissen ist eine inhärente menschliche Fähigkeit, im Beruf
bleibt dieser Umgang jedoch oft implizit. Kompetenter Umgang mit Nichtwissen als eine kulturelle
Herausforderung könnte durch Veränderungen in der Einstellung, im Verhalten und in
institutionellen Strukturen befördert werden. Einige Prinzipien von kompetentem Umgang mit
Nichtwissen werden für die weitere Erforschung zusammengestellt. Durch Hinzunahme von
Variablen für Umgang und Evaluierung können entscheidungsbezogene Unbekannte mit der
vorgeschlagenen Heuristik so kartiert und kodiert werden, dass sie das „schriftlose“ Nichtwissen
abbilden. Berufstätige, die im Umgang mit Nichtwissen alphabetisiert sind, könnten ihre
entscheidungsrelevanten Unbekannten verstehen, erklären und strategisch nutzen.
Schlagworte: Nichtwissen, Entscheidungen, Heuristiken, Ökonik, Post-normale Wissenschaft,
Panarchy
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
5
III. Table of Contents
I. Abstract ........................................................................................................................................... 3
II. Zusammenfassung (German Abstract) ............................................................................................ 4
III. Table of Contents ............................................................................................................................ 5
IV. Table of Figures ............................................................................................................................... 8
V. Acknowledgements ....................................................................................................................... 10
1. Introduction ......................................................................................................................... 11
1.1 Rationale and the Research Process ............................................................................................. 13
1.2 Econical Approach ......................................................................................................................... 14
1.3 Objective, Research Questions and Design ................................................................................... 16
2. Exploitation Phase: Theoretical Part ................................................................... 17
2.1 Use of Literature to Clarify Essential Concepts ............................................................................. 19
2.1.1 Thinking ................................................................................................................................. 19
2.1.2 Decisions ................................................................................................................................ 21
2.1.3 Heuristics ............................................................................................................................... 23
2.1.4 Knowledge ............................................................................................................................. 27
2.1.5 Non-knowledge ..................................................................................................................... 28
2.1.6 Non-knowledge Literacy ........................................................................................................ 30
2.2 The Post-normal Science Diagram................................................................................................. 33
2.3 Exploring the Non-Knowledge Map for Decisions ......................................................................... 33
2.4 Results and Hypotheses from Theoretical Part ............................................................................. 35
3. Conservation Phase: Empirical Part ..................................................................... 36
3.1 Interviewees .................................................................................................................................. 36
3.2 Methods ........................................................................................................................................ 37
3.2.1 Questionnaire ........................................................................................................................ 37
3.2.2 Interviews .............................................................................................................................. 42
3.2.3 Qualitative Analysis ............................................................................................................... 43
3.2.4 Quantitative Analysis ............................................................................................................. 45
3.2.5 (Non-)Availability of Data ...................................................................................................... 46
3.3 Results ........................................................................................................................................... 46
3.3.1 Collected Unknowns and the Non-knowledge Map for Decisions ........................................ 46
3.3.2 How Did You Take the Decision? ........................................................................................... 54
3.3.3 Non-knowledge Literacy ........................................................................................................ 55
3.3.4 Post-normal Science Diagram Inspired Questions ................................................................ 56
3.3.5 Exemplary Interview .............................................................................................................. 58
3.3.6 Exemplary Decisions .............................................................................................................. 60
3.3.7 Discursive Interviews ............................................................................................................. 61
3.3.8 Remainder of Information from Interviews .......................................................................... 65
3.3.9 Statistical Analysis ................................................................................................................. 67
3.4 Summary of Results from Empirical Part....................................................................................... 70
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
6
4 Release Phase: Discussion .......................................................................................... 70
4.1 Discussion of Theoretical Part ....................................................................................................... 70
4.1.1 Literature ............................................................................................................................... 70
4.1.2 Exploring the Non-knowledge Map (Compass and Table Format) ....................................... 71
4.1.3 Discussion of Adapted Post-normal Science Diagram ........................................................... 71
4.2 Discussion of Empirical Part .......................................................................................................... 71
4.2.1 Methods ................................................................................................................................ 72
4.2.2 Results ................................................................................................................................... 80
4.3 Mapped Unknowns ....................................................................................................................... 86
4.4 Non-knowledge Map ..................................................................................................................... 88
4.4.1 Temporality ........................................................................................................................... 90
4.4.2 Ambiguity .............................................................................................................................. 90
4.4.3 Knowability ............................................................................................................................ 90
4.4.4 Solution Relevance ................................................................................................................ 91
4.4.5 Threat Potential ..................................................................................................................... 91
4.4.6 Intentionality ......................................................................................................................... 92
4.4.7 Recognition ............................................................................................................................ 93
4.5 Non-knowledge Literacy ................................................................................................................ 93
5 Reorganisation Phase: Post Analysis Review ................................................ 95
5.1 Methods for Reorganisation ......................................................................................................... 95
5.1.1 Material ................................................................................................................................. 95
5.1.2 Reflections and Definitions ................................................................................................... 96
5.2 Results ......................................................................................................................................... 100
5.2.1 Coded Exemplary Interview ................................................................................................ 100
5.2.2 Coded Interview Unknowns ................................................................................................ 101
5.2.3 Non-knowledge Literacy ...................................................................................................... 103
5.3 Concluding Discussion ................................................................................................................. 104
6 Conclusion ........................................................................................................................... 106
6.1 Answers to Research Questions .................................................................................................. 106
6.1.1 What is non-knowledge literacy? ........................................................................................ 106
6.1.2 How can the non-knowledge map be applied to and improved for decisions? ................. 106
6.1.3 How volatile is the working environment of Global Change Managers? ............................ 107
6.1.4 Which heuristics do Global Change Managers use for work decisions? ............................. 107
6.1.5 Which unknowns are common in work decisions? ............................................................. 107
6.1.6 Are Global Change Managers able to manage non-knowledge? ........................................ 107
6.1.7 Are Unknowns Impeding Effective Decision Making? ......................................................... 107
6.2 Conclusions on the Decision-centred Exploration of Non-knowledge ........................................ 107
6.3 Recommendations for Decision Makers ..................................................................................... 109
6.4 Outlook and Final Reflections ..................................................................................................... 109
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
7
VI. Bibliography ................................................................................................................................. 111
VII. Annex A........................................................................................................................................ 117
1. Board Game Map ........................................................................................................................ 117
2. About Global Change Management (MSc) .................................................................................. 118
3. Trial Run ....................................................................................................................................... 122
4. Exploring (the) Non-knowledge (map) II ..................................................................................... 125
VIII. Digital Annex B ............................................................................................................................ 130
1. Non-knowledge Literacy Quotes ................................................................................................. 130
2. Questionnaire Versions 0.0-1.4 ................................................................................................... 130
3. Terms for Coding (how did you decide) ...................................................................................... 130
4. Clustering for Coding (how did you decide) ................................................................................ 130
5. PCA Results .................................................................................................................................. 130
6. PCA Data ...................................................................................................................................... 130
7. Coding (Keys and Abbreviations used for PCA) ........................................................................... 130
8. Clustering of (Mapped) (Non-)knowledge................................................................................... 130
9. Exposé.......................................................................................................................................... 130
10. Development of Questionnaire ................................................................................................... 130
IX. Declaration of Independent Work on Master Thesis .................................................................. 131
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
8
IV. Table of Figures
Figure 1 shows a simple sketch representing how a Global Change Manager (MSc) might explore information
and the according knowledge and non-knowledge to come to a decision. 11
Figure 2 shows the panarchy (Gunderson and Holling, 2002) and which part of this thesis embodies which
phase on the panarchy. The four phases that can be distinguished are exploitation, conservation, release
and reorganization. They are located along the axes of potential and connectedness. The theoretical part
represents the exploitation phase, the empirical part conserves it and has the highest potential and
connectedness, the discussion then represents the release and the Post Analysis Review the
reorganisation phase. This thesis then ends with a new concept which is open for exploitation. 14
Figure 3 The non-knowledge map (Ibisch and Hobson, 2012a) shows nine dimensions of non-knowledge. Those
dimensions are expressed as axes or gradients between different forms of non-knowledge, e.g. from
knowable to unknowable or intentional to unintentional. It was hypothesised that the stronger extremes
of the gradients in the dark part of the map are more relevant to sustainability than those in the light
part. Concrete types of non-knowledge are found in bubbles. This initial conceptualisation of the non-
knowledge map is also referred to as the compass version. 18
Figure 4 contrasts system 1 and system 2 abilities as presented in Kahneman (2011) 20
Figure 5: the quote illustrates "counter-rational" behaviour of rats which is functional in a natural environment.
24
Figure 6 illustrates the concept of antifragility. The quote is taken from part II of the preface of the book
Antifragile Things That Gain From Disorder (Taleb, 2012). 25
Figure 7 introduces some heuristics and biases which are referred to in this thesis. 26
Figure 8 distinguishes knowledge and non-knowledge at meta-level. The first found use of this distinction was
by Rumsfeld (2002) whose speech became popular for the concept. 28
Figure 9 The post-normal science diagram (Funtowicz and Ravetz, 2003) provides a framework to map systems
uncertainty against decision stakes. It postulates that applied science can be used when both are low,
professional consultancy when at least one is medium and post-normal science when systems uncertainty
or decision stakes are high. 33
Figure 10 shows the simple table format non-knowledge map. For evaluation numbers from -2 to 2 have been
assigned. The numbers only express that the two ends of each dimension are bipolar. 34
Figure 11 shows the adapted table format non-knowledge map as used in questionnaire version 1.4. 35
Figure 12 shows the distribution of 40 unknowns according to the gradients of the dimensions. The colours
represent the code used throughout this thesis (compare Figure 11). The specification to the right gives
the concrete terms for each presented colour and is probably more intelligible than the -2 to 2 coding. 48
Figure 13 shows the result of the post-normal science diagram inspired questions. The green bars show the
distribution among low, medium and high for the respective question. The blue bars give the average of
the preceding green bars and answer the respective overarching question. 56
Figure 14 shows all decisions on the post-normal science diagram. Green dots or background means low,
yellow medium and red high idiosyncrasy. The colours could be extrapolated globe-like into space. The
red line (two dimensional only) shows that decision stakes vary less among decisions than the volatility of
the environment. 57
Figure 15 shows the results of a PCA with the data set Cleaned Values II that has a variance of only 12%. Dots
are given numbers, the first digit is the interview, the second the decision and the third the unknown:
5.12 is the data set for the second unknown of the first decision in the fifth interview. Groups for how to
(dark blue), if to (black) and what exactly (light blue) decisions are displayed. Biplots are too close to be
sensibly evaluated. Decision pairs are usually close. 68
Figure 16 shows the result of a PCA with the seven dimensions of non-knowledge for 40 maps. The first and
second components explain 26% and 22% of the variables respectively. This PCA suggests that (1)
Knowability (NM_Kno) and Ambiguity (NM_Amb) are independent from relevance (NM_Rel),
intentionality (NM_Int) and recognition (NM_Rec), (2) unknowable (NM_Kno) and ambiguous (NM_Amb)
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
9
non-knowledge tends to have a high threat potential (NM_Thr) and tends to be related to the future
(NM_Tem) and (3) unintentional (NM_Int) non-knowledge seems to be relevant (NM_Rel) and hence
intentional non-knowledge seems to be irrelevant. Dots represent maps but no additional pattern could
be observed so that no numbers are displayed. 69
Figure 17 shows the graphical representation of the unknown “I do not know what will change”. It maps
knowledge and non-knowledge as part of one continuum against time. The perspective of the interviewee
is blue, a more general societal perspective is black and the handling is green. The decision point, the
interview day and the moment of manifestation are indicated. Manifestation is when the unknown turns
into something known, it is the moment of revelation. For this unknown, manifestation is the moment for
which the unknown is relevant. Before the interview, the personal recognition of the unknown was
limited (dashed line). This generic unknown was recognised in the interview. It is assumed to remain
recognised until the moment of the coaching (manifestation). Simultaneously, the change might (dashed)
already have happened or be known by others (existence or societal recognition). For this unknown,
ambiguity overlays all lines. The interviewee perceived the unknown change as potentially benefiting. The
proposed handling was scenario thinking, adaptive management and trust in own capacities. It bridges
the gap to the point of relevance. 84
Figure 18 builds on the meta-distinction of knowledge and non-knowledge (compare 2.1.5). Meta-unknowns
are blindspots and meta-knowns can either be knowns or unknowns. In the figure, the blindspot is largest
as it is continuously fed by oblivion and the infinite unknown that Ibisch and Hobson (2012b) refer to as
dark matter. The unknown is fed by oblivion of knowns. The (non-)knowledge would be held by an
individual, a group or society. The black processes should be considered for the question at hand. A
blindspot can either become an unknown by recognition, or it can become a known by manifestation. An
unknown can become a known by manifestation. Oblivion can reverse these processes. Blindspots
created by oblivion can usually be reversed easily. From now on, terminologically, a blindspot is
distinguished from an unknown. 97
Figure 19: the tables provide the key to code an unknown. The points in time have to be ordered for the
specific situation. A class of non-knowledge describes R and a class of knowledge M. The (non-)knowledge
classes are ordered according to scale from blindspot (low) to clearly emerged (high). Handling and
evaluation are not specified in detail. 100
Figure 20 shows the coded results for all 40 interview unknowns, five scenario unknowns (sc.) and the one
interview decision that did not have any unknown as it was taken by a directive (none). They are ordered
according to timing. Blocks of same colour have the same timing. Unknown and known provide the
gradient of (non-)knowledge at X and M respectively. 102
Figure 21 gives a do-it-yourself instruction on how to heuristically map an unknown for a past or future
decision. 103
Figure 22 shows the non-knowledge map transferred into a board game. Additionally the user needs cards with
both ends of the dimensions (front and back) which have to be placed on the schedule. A draft version of
the game instructions is written but not published. 117
Figure 23 First-year GCM students from 2006 to 2014, figure adapted from Günther-Dieng et al. (2014) 119
Figure 24 GCM students from 2006-2014, figure adapted from Günther-Dieng et al. (2014) 120
Figure 25: All mapped non-knowledge from trial run interview decisions was recognized, related to the future
or present and had a threat potential. For some the geographical or social distribution dimensions did not
apply. There are more bars in the black “more sustainability relevant” part of the map. Figure 10 provides
the key for the gradients. 123
Figure 26: The decisions cluster strongly in the second and partly in the third realm of the diagram. In three of
four cases the first and second decision are placed in a similar place on the diagram and apply the same
decision "method". 124
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
10
V. Acknowledgements
In the past six month, few conversations passed without me referencing the unknown. I saw and see
non-knowledge everywhere. I would like to extend my deepest thanks to all those who allow me to
feel this deep trust I feel.
I would like to thank my supervisors Prof Dr Pierre Ibisch and Dr Peter Hobson, for all the interesting
work you published, the time you took and the thoughts, concepts, knowledge and non-knowledge
you shared with me.
I would like to thank my interviewees for the time you took, the nice and interesting talks we had
and the decisions and unknowns you shared.
I would like to thank the authors I read for all the interesting knowledge and non-knowledge you
shared.
I would like to thank Christoph Nowicki for the support in finding interviewees.
I would like to thank my friends. I extend my special thanks to those friends who have helped me to
get more knowledge out of this thesis. Thank you, Thorleif, Alex, Elena, Daniel, Christina and Mirko,
for proofreading and commenting on various parts of this thesis. Thank you, Ingo and Julia, for
helping me to structure my conceptual overload. Thank you to Sebo, Sophie, Leon, and the others
who have mapped their unknowns. Thank you, Bine and Davide. And thanks for those nice on- and
off-topic lunches we had.
I would like to thank my aunt, Carolin, for the first mapped unknown, my dad for the intentional
unknowns mapped, the suggested literature and the shared interest and my mom and Peter for
playing the non-knowledge board game.
Thanks to all those I do not know I could to be thankful to.
And, thank you mom, for how you taught me. And everything else.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
11
1. Introduction
In times of global change, knowledge is increasing exponentially. Practitioners, such as Global Change
Managers, have to deal with an increasing amount of incoming information as well as with a large
amount of knowable information and immense non-knowledge. All this influences their decision-
making at work (see Figure 1). Heuristics are useful and frequently used tools to navigate the
information jungle – and to facilitate decision making. Non-knowledge literacy (Ibisch and Hobson,
2012a) is a concept that allows competent handling of the unknown. A multidimensional analysis of
unknowns from decisions taken by Global Change Managers is carried out in this thesis. The common
handling of those unknowns is compared to principles of non-knowledge literacy from literature. The
non-knowledge map for decisions is developed into a simple heuristic that allows the further
exploration and creation of non-knowledge literacy.
Figure 1 shows a simple sketch representing how a Global Change Manager (MSc) might explore information and the
according knowledge and non-knowledge to come to a decision.
Uncertainty is a non-knowledge form that is frequently discussed in literature. However, few
attempts have been made to understand these uncertainties in depths. The more generic and value-
neutral term non-knowledge includes uncertainties, but also other types of non-knowledge. The non-
knowledge map (Ibisch and Hobson, 2012a) provides a systematic starting point to the exploration of
non-knowledge. Hobson and Ibisch (2012) propose heuristics as uncertainty-accepting ways to make
decisions.
Heuristics are basic rules that demonstrated resilience under change. Heuristics allow competent
acting in uncertain environments. Adaptive minds functioning with heuristics work best in a world
with new challenges evolving continuously. The following excerpt illustrates the paradox:
“Humans have evolved with heuristic tendencies - by nature we are Homo heuristicus rather
than Homo algorithmicus - in order to better cope with the uncertain and unpredictable
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
12
world. These traits enable our species to “muddle” through periods of change and this
resembles the way that complex ‘knowledgeable’ ecosystems adapt to change by using some
of the potentially available knowledge, but not all of it. Paradoxically, as it has turned out,
Homo heuristicus, has become so successful that as a social super-organism, our species has
developed a cultural knowledge and wisdom that has promoted diagnostic traits, namely,
algorithmic thinking and planning. The emergence of this evidence-based science has not
always yielded positive benefits and may have even contributed to the kinds of problems now
witnessed in the ambitious attempts to harness technology for the purpose of macro-
managing global ecosystems. Human capabilities to create enormous amounts of knowledge
about any part of the complex global system have created unfortunate negative feedback
problems that are now having undesirable consequences for human wellbeing. There is a case
for the argument that knowledge overload encourages overcautious tendencies that
constrain brave decision-making. Other emerging patterns include problems of using
predictive models and evidence in decision-making policy that are misguided and
inappropriate in a dynamically changing complex system. Thus, in times of rapid global
change, the challenge now is to seek more sophisticated forms of heuristics that work at
multiple scales from the smallest at the individual level to the biggest at national, regional or
even global levels –a revolutionary challenge for science. It would have to guide us through a
new learning experience of working competently with black boxes that represent imperfect
and incomplete knowledge.” (Hobson and Ibisch, 2012)
• Heuristics allow humans to cope with uncertainties
• Ecosystems do not use all available knowledge
• Decision making might be hampered by knowledge overload or static knowledge
• Science could create overarching heuristics for times of rapid global change
• We could learn to work with black boxes of imperfect and incomplete knowledge
New scientific practices which are inclusive towards non-knowledge and look beyond scientific
evidence are for example a post-normal science perspective that acknowledges cognitive limitations
or an econical approach (Freudenberger et al., 2010; Ibisch and Hobson, 2014). Both are used in this
thesis.
Sometimes the unknown can be known, but often the unknown remains unknown - unknown at least
for the moment. Knowledge about the patterns of the unknown can be accumulated. This might
increase the understanding of unknowns and should allow their responsible handling. This thesis
provides a cautious attempt to find such patterns in closely defined unknowns from decisions in
professional contexts.
Based on the idea that non-knowledge needs a subject and a context, non-knowledge of individual
decision makers is investigated in this thesis. Non-knowledge is assessed by the non-knowledge map
which was designed as a systematic tool for decision unknowns. Decisions provide situations for the
application of non-knowledge. Decisions are about attaching value. Application of the map had to
happen by interviews. In personal communication (Ibisch, 2015) the author of the non-knowledge
map said it had not been used before.
Management implies decision making. Global Change Managers, the group of interviewees, present
a comparably heterogeneous group of what might be called young professionals. There are currently
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
13
six to seven generations of the study programme´s alumni. This cautious convenience sample was
also taken because the research design is experimental and it was unknown if it produces useful
results.
Complex systems can be understood in their attributes and functionality but their exact
manifestation remains unknown. The unknown is more important to sustainability than any knowns
(Ibisch and Hobson, 2012a). Taleb (2008) formulates “I will never get to know the unknown since, by
definition, it is unknown. However, I can always guess how it might affect me, and I should base my
decisions around that.” The term ‘non-knowledge literacy’ is proposed by Ibisch and Hobson (2012a).
Non-knowledge literacy claims that the processes and results from unpredictable change need to be
equally considered in education and decision making as known phenomena (Ibisch and Hobson,
2012a). Non-knowledge literacy would promote a strategic and operational inclusion of non-
knowledge and what can be known about it(Ibisch and Hobson, 2012a).
1.1 Rationale and the Research Process
The author´s study programme postulates global change. This change is accelerating (Steffen et al.,
2011). One of its manifestations is the knowledge explosion. This knowledge explosion often does
not reap additional benefits for the practitioner (Ibisch, Vega and Herrmann, 2010). The practitioner
and manager has to take decisions, many of those by integrating future scenarios (Ibisch and Hobson,
2012b). One potential group of practitioners are alumni of the study programme itself. The equally
titled book “Global Change Management: Knowledge Gaps, Blindspots and Unknowables” (Ibisch,
Geiger and Cybulla, 2012) provides a hook to this self-referential and heuristically set investigative
circle.
Largely undirected acquisition of knowledge during various weeks led to the conceptualisation of the
rationale which then used a simple heuristic "write something related to the study programme". The
book “Global Change Management: Knowledge Gaps, Blindspots and Unknowables” (Ibisch, Geiger
and Cybulla, 2012) was a main source and provided the initial hypotheses and concepts: The non-
knowledge map, the term non-knowledge literacy and that heuristics might be worth investigation.
The initial title then spelled: “Non-knowledge Literacy – Working Heuristics of Global Change
Managers”. The exposé from April 2015 can be found in annex VIII.9.
Some aspects were adapted during the research process. In interviews, the concept was
implemented. The whole research process was accompanied by a manuscript. This manuscript
combined elements of chaos and order, it remained chaOrdic (Hobson and Ibisch, 2012). Its content
was hard to grasp – even for the author. This forced a rewriting of large parts of the manuscript.
Evolved structures of scientific works were used to make the findings explicit and clear. This a
posteriori application of structure allowed the emergence of most results. Within two weeks, the title
was changed twice. Towards the last week, the title spelled “Exploring Non-knowledge Literacy – A
Multidimensional Analysis of Unknowns at Work”.
Some of the chaOrdic design of the manuscript is retained. It allows the intuitive reader to get a
different understanding of the text and teases the critical reader. Those readers that can overcome
the chaOrdic structure might be nudged into a new understanding of communicated uncertainty and
the panarchic cycle of non-knowledge.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
14
The main sections of this thesis are structured along the four phases of the panarchy (Figure 2). The
theoretical part exploits terms and concepts. The generated knowledge is conserved in the empirical
part. In the discussion, concepts are released. In the post analysis review, they are reorganised into a
non-knowledge mapping heuristic for decisions: a new tool open for exploitation. The thesis is
framed by this introduction and a conclusion. The use of the panarchy to structure this work is one
expression of the econical approach.
Figure 2 shows the panarchy (Gunderson and Holling, 2002) and which part of this thesis embodies which phase on the
panarchy. The four phases that can be distinguished are exploitation, conservation, release and reorganization. They are
located along the axes of potential and connectedness. The theoretical part represents the exploitation phase, the
empirical part conserves it and has the highest potential and connectedness, the discussion then represents the release
and the Post Analysis Review the reorganisation phase. This thesis then ends with a new concept which is open for
exploitation.
1.2 Econical Approach
Today´s evidence-based approaches to science are questioned and non-knowledge based
approaches offered as an alternative (Ibisch, Vega and Herrmann, 2010). Econics (see box) have
investigated how knowledge and non-knowledge are managed in nature (Hobson and Ibisch, 2012).
Some of the principles found are applied in this thesis.
Science is a natural resource-dependent socio-economic system. Science has always explored the
unknown. Science is a systematic generator and manager of knowledge. In this case, science shall be
used as a gateway to explore sustainable development under global change. This is why it is
postulated that the research design and questionnaire at hand were generated by an econical
“Econics is a transdisciplinary approach to studying the dynamics and functioning of (complex and
holarchically nested) ecological systems with the aim of deriving management solutions for natural
resource-dependent socio-
economic systems as a gateway towards sustainable development
under global change.” (Hobson and Ibisch, 2012)
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
15
approach to science: a transdisciplinary, metasystemic and non-knowledge based approach which
makes heuristics, redundancies and omission as well as uncertainties and buffers explicit.
The panarchy, along which this thesis is structured, is “an example of econics par excellence”
(Hobson and Ibisch, 2012).
The concept of econics is comparably young (introduced by Althaus in 2007) and still evolving
(Hobson and Ibisch, 2012). Examples of econical approaches to the study of forests (Hobson and
Ibisch, 2013) and (non-)knowledge in nature (Hobson and Ibisch, 2012) exist. Econics refer to
efficient and functioning systems that have resource and spatial limits and are subject to
disturbances and changes in the environment (Hobson and Ibisch, 2012). A “close-to-nature” social
knowledge framework based on the characteristics of non-knowledge in nature and culture is
suggested as a case for econics (Ibisch and Hobson, 2012b).
A scientific work has resource and spatial limits, the results were not found as linear as presented,
they were subject to disturbances and changes, e.g. due to new scientific information or surprises in
the results.
Econical guidelines for knowledge managers (this includes scientists) proposed by Hobson and Ibisch
(2012) include inter alia an analysis of non-knowledge relevant for sustainability, a typology (or
register) for forms of non-knowledge, cautiousness in postulating evidence-based decision making
and acknowledging heuristics for uncertain situations. The topic at hand is uncertain and can only be
approached by heuristics. Heuristics are econical (Ibisch and Hobson, 2012a). Metasystemic
management mimicks self-regulatory processes of complex (eco)systems and is an example of an
econical approach (Ibisch, Vega and Herrmann, 2010). Transdisciplinarity imitates patterns of
information and networks in nature as it retains a broad understanding and manifold
interconnections (Ibisch and Hobson, 2012a).
Global Change Management is designed as a transdiscipline (e.g. Ibisch, 2010). This means that
transdisciplinarity is understood as the dissolution of boundaries between disciplines and the
integration of non-scientists and context-loaded research environments. It transcends scientific
disciplines. In this understanding, transdisciplinarity requires including knowledge from traditional
disciplines in a metasystemical way. Such a metasystemic integration is often non-knowledge based,
as some disciplinary knowledge is only useful for those trained in the discipline whereas other results
can be used and integrated across disciplinary boundaries.
Such transdisciplinarity is not seen as a lack of complementary knowledge (about routines, best
practices, common approaches, mistakes that have been made and should be avoided, etc.) but as a
chance. Ignorance can cause positive change (Gigerenzer, 2007a). It is a free and heuristic way of
approaching the question with explicit ignorance and common sense. This might lead to mistakes.
But a known approach is not exempt from previously made mistakes either. Science is never
objective (Ibisch and Hobson, 2012a).
If disciplines were to be assigned, these would most certainly include psychology and philosophy.
Interviews and reflections are used in this thesis. Both are widely agreed scientific methods which
are used across disciplines and gain importance (Eisenhardt und Graebner 2007). Their individual
design and subsequent analysis is often strongly determined by best practices within the discipline. A
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
16
posteriori, attempts have been made to justify the applied methods and match them with common
methods as a solid basis if this research is to be furthered.
Heuristics are metasystemic rules that are explicitly based on incomplete information. Heuristics are
non-knowledge based. Simple guiding rules were used in research and interview design: convenient
literature, a convenient sample and simple categories. Metasystemics allowed omission. Not all
collected data were explored –nature does not do that. Some generated information (e.g. form
recordings) were not analysed in depth – they were intentionally not transferred into knowledge.
The statistical analysis was only carried out for a few data sets – this is omission. Some “redundant”
data was left aside. Redundancies are an investment (Taleb, 2012) or an insurance and buffer.
Collecting more data than those that would be used is precautionary.
1.3 Objective, Research Questions and Design
It is the objective of this thesis to explore non-knowledge to gain a better understanding of the term
non-knowledge literacy in global change related decision making.
This shall give first impressions of the types of non-knowledge that are relevant for global change
related decisions, how these types can be clustered, and how the non-knowledge map can be applied
methodologically. Furthermore the results shall enrich the vocabulary to talk about uncertainties in
decision making and feed back into the advancement of the Global Change Management study
programme (GCM). The method shall also be tested for potential further application.
It should be the aim of this thesis to find a suitable, memorisable and easy-to-use non-knowledge
map for decision situations. The current non-knowledge map (compare Figure 3) is more like a
compass that does not indicate how its nine dimensions are linked. A non-knowledge map for
decisions might then build upon strong links between those dimensions.
Hypothesis-driven research requires methods that aim at falsifying the hypothesis. This often goes
wrong, as people are prone to confirmation bias and tend to design methods that confirm their
hypothesis (compare e.g. Kahneman (2011) and Taleb (2008)). As non-knowledge literacy is a rather
new and open concept that is not strictly defined or definable, a hypothesis does not apply. Rather a
statement is made that is to be explored by the methods described below:
A competent understanding of non-knowledge is a key element of non-knowledge literacy.
Together with a bias-conscious use of heuristics it can facilitate decision making under
uncertainty.
The underlying postulate of this thesis is that non-knowledge literate individuals are happier and
take better decisions and that, on an upscaled institutional level, those decisions lead to more
effective implementation.
Conceptual work with the non-knowledge map is at the heart of this scientific endeavour. Literature
and an interview series with subsequent statistical evaluation broaden the perspective. The non-
knowledge map (compass) (2.3), the post-normal science perspective (2.2), six key terms (2.1) and
information from 21 interviews (3.1) provided material for this thesis.
The methods to explore elements and principles of non-knowledge literacy are based on an econical
approach. The literature is closely defined and moves from thinking over decisions to heuristics and
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
17
from knowledge and non-knowledge to non-knowledge literacy. A series of 30 minute interview calls
with 20 Global Change Management alumni is carried out on two decisions from their professional
life. The questionnaire assesses a wide range of variables concerning the decisions, their
environment and the associated (un)knowns. Two specific unknowns are mapped with the
interviewees. A multivariate analysis is carried out for the interview data. Before and after the
interviews, the map is explored and improved for decision situations by tinkering. Tinkering, as used
by Taleb (2008) would be a form of bottom-up undirected trial and error. Methods, results and the
following research questions are discussed:
• What is non-knowledge literacy?
• How can the non-knowledge map be applied to and improved for decisions?
• How volatile is the working environment of Global Change Managers?
• Which heuristics do Global Change Managers use for work decisions?
• Which unknowns are common in work decisions of Global Change Managers?
• Are Global Change Managers able to manage non-knowledge?
o “The ability to manage non-knowledge will ultimately define a good global change
manager.” (Ibisch and Hobson, 2012a)
• Are unknowns impeding effective decision making?
o “The apparent vastness of unknown phenomena is to many the very impediment to
effective decision-making and problem-solving.” (Ibisch and Hobson, 2012a)
o “little or no knowledge is the source of vulnerability and insecurity that can and do
lead to masking strategies that then lead on to most heuristic decisions being made –
the stuff of nightmares in the medical and scientific professions. BUT is where so
much of the real world is at!!!”, personal communication (Hobson, 2015)
2. Exploitation Phase: Theoretical Part
The theoretical part explores the topics at hand using literature and (thought) experiments. It then
forwards results and hypotheses to the empirical part for further investigation.
The topic of interest, non-knowledge and decisions, are linked by the following to statements
• Just as knowledge, non-knowledge needs a subject.
• Decisions are concrete occasions for the occurrence of unknowns.
An individual can hold non-knowledge and so be a “subject”: The individual does not know. There are
other potential subjects, such as groups or nature. There is non-cognitive knowledge in humans and
non-human cognitive knowledge (Hobson and Ibisch, 2012) - those comprise non-knowledge. Hence
it can be said that there must be a system which holds the non-knowledge. In this thesis, these
systems are individual human decision makers.
Literature was used to clarify the six key terms and concepts (see box). The exploration of non-
knowledge required an understanding of knowledge. One conceptual tool for handling non-
knowledge is non-knowledge literacy. Thinking is one basis for decisions. Heuristics are a concept
that unites various decision tools.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
18
Exploration of the non-knowledge map (see Figure 3) for decision situations generated a first
understanding of the map and necessary adaptations for decision situations. The non-knowledge
map is a tool that promises a systematic exploration of the unknown. It is the core concept which is
explored by tinkering, reflection and trial.
Figure 3 The non-knowledge map (Ibisch and Hobson, 2012a) shows nine dimensions of non-knowledge. Those
dimensions are expressed as axes or gradients between different forms of non-knowledge, e.g. from knowable to
unknowable or intentional to unintentional. It was hypothesised that the stronger extremes of the gradients in the dark
part of the map are more relevant to sustainability than those in the light part. Concrete types of non-knowledge are
found in bubbles. This initial conceptualisation of the non-knowledge map is also referred to as the compass version.
The non-knowledge map in the initial compass format (Figure 3) is a tool that allows a systematic
exploration of the unknown. It is the core concept which is explored, adapted, applied and modified
in this thesis.
Nine dimensions are suggested: intentionality, ambiguity, geography, social distribution, knowability,
solution relevance, threat potential, temporality and recognition. They do not appear to be sensibly
ordered. The terms at either ends, the poles, appear realistic and the types of non-knowledge
Basis Topic of interest Conceptual tool
Knowledge <- Non-knowledge -> Non-knowledge literacy
Thinking <- Decisions -> Heuristics
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
19
familiar. An in depth description of each dimension is given by Ibisch and Hobson (2012a). Excerpts of
those descriptions and a first understanding from a decision perspective with closely defined
unknowns can be found in 3.3.1.
2.1 Use of Literature to Clarify Essential Concepts
Not only in times of global change and the knowledge explosion it is impossible to access all relevant
knowledge (Ibisch and Hobson, 2012a). It is impossible to consult all potentially relevant literature.
Some reductionism and metasystemic management has to be applied. This is econical.
The focus was deliberately and heuristically set on overarching and stock-taking or on popular
scientific and essayist literature for non-knowledge. For decisions, mainly popular scientific works
were consulted. The use of popular scientific works from renowned scientists allowed access to
results from disciplinary studies. It is assumed that popular scientific works contain ‘what has trickled
down from science to the public’. Consulting a higher share of primary literature was consciously
prevented to stay within the (temporal) circumstances and explorative nature of this thesis. Relying
on secondary literature is metasystemic.
Literature was consulted throughout the research process. In the beginning, literature was screened
and scanned. In-depth reading and analysis was done during tinkering, trial run and questionnaire
design. Some complementary readings and rereading later in the process took place.
Initial authors, as defined in the exposé (annex VIII.9), were Gigerenzer (heuristics), Ibisch and
Hobson (non-knowledge), Kahnemann and Tversky (heuristics and bias) as well as Taleb
(uncertainty). Additionally, Klein (1999) was consulted on expert decisions under uncertainty.
For a comprehensive and general introduction into decision making, Taleb (2008) recommends Baron
(2000), hence the updated Baron (2008) was roughly scanned and used in case of doubt. A short
expedition was done into war literature (Giles, 1910; von Clausewitz, 1832). Any further literature
used was based on the bibliographies of the above mentioned works (e.g. essays published in Vitek
and Jackson (2008b) provided useful insights) or searched to close recognized and solution-relevant
knowledge gaps (e.g. Barker et al. (2007) compile and discuss research on dealing with uncertainties
in decisions).
The literature mainly used (Gigerenzer, 2007b; Taleb, 2008; Kahneman, 2011; Taleb, 2012) are
popular scientific monographs written by single scientists that have coined a field and are well
known. These monographs provided access to scientific results and gave extensive explications of the
topic, underlying motivations and the author´s mind-set. They also pointed to scientific literature.
In the following, terms and concepts used in this thesis are clarified. Those include thinking as a basis
for decisions, decisions and heuristics as a decision making tool. The second line of enquiry goes from
knowledge over non-knowledge to non-knowledge literacy.
2.1.1
Thinking
Thinking is the basis of decisions and can be fast or slow (Kahneman, 2011). Baron (2008) defines
thinking as “what we do when we do not know what to choose, desire, or believe” and says it begins
with doubt, needs transfer and is often required in new situations. It is furthermore understood as an
overlapping search and inference process for which different people have different thinking
strategies (heuristics) that can be learned (Baron, 2008).
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
20
Kahneman´s (2011) understanding of fast and intuitive system 1 thinking and effortful slow system 2
thinking is used in this thesis. System 1 produces most thoughts and system 2 is used to solve difficult
problems. System 2 usually checks and endorses what system 1 proposed. The following quotations
illustrate their characteristics:
System 1
System 2
automatically and quickly, with little or no effort
and no sense of voluntary control
Allocates attention to the effortful mental
activities that demand it
Continuously running, cannot be turned off,
difficult not to do more than System 2 charges it
to do
It is lazy. It is mobilized when system 1 does not
offer an answer.
Continuously generates suggestions for System
2: impressions, intui
tions, intentions, and
feelings
If endorsed by System 2, impressions and
intuitions turn into beliefs, and impul
ses turn
into voluntary actions
Assigns causality and subsequent interpretation
and integration of information
Can follow rules, compare objects on several
attributes, and make deliberate choices between
options
Easily thinks associatively and metaphorically
More detailed and specific processing
Is often not noticed by the conscious self
It´s activity can be physically observed (e.g. by a
dilated pupil)
Many things happen simultaneously although it
cannot deal with multiple d
istinct topics at the
same time
Limited capacity for interfering activities (“you
cannot calculate 17*24 while making a left turn
in dense traffic)
Can be primed (e.g. a forced smile makes you
interpret comics funnier)
Can be mobilized by cognitive strain and
weakened e.g. with tiredness or alcohol
Cannot deal with statistics, is easily biased
Can do complex computations
Intuitive thoughts can be caused by experience
(expert intuition) or the use of heuristics or by
entirely automatic mental processes such as
memory and perception
It can programme automatic System 1 functions
of e.g. attention and memory
It has learned skills, such as reading and
understanding of social situations
It monitors and controls. It is in charge of self-
control and can overcome intuitions and
impulses from System 1
It is more influential than your experience tells
you, and it is the secret author of many of the
choices and judgments you make
Often associated with the subjective experience
of agency, choice, and concentration
It is radically insensitive to both the quality and
the quantity of the information that gives rise to
impressions and intuitions
It is more of an apologist for the emotions of
System 1 than a critic of those emotions—an
endorser rather than an enforcer. Its search for
information and arguments is mostly constrained
to informa
tion that is consistent with existing
beliefs, not with an intention to examine them.
Figure 4 contrasts system 1 and system 2 abilities as presented in Kahneman (2011)
A frequent problem encountered in thinking is what Kahneman (2011) describes as “an active,
coherence-seeking System 1 suggests solutions to an undemanding System 2.“. A complete reliance
on system 2 would be too tedious, impractical and not necessarily good. System 2 is too slow and
inefficient for routine decisions and continuously questions the own thinking. A compromise is
suggested: “learn to recognize situations in which mistakes are likely and try harder to avoid
significant mistakes when the stakes are high.” (Kahneman, 2011)
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
21
According to Baron (2008), thinking extends into decision making, e.g. if it is inventive or creative,
such as the creation of scientific theories, or planning. Such thinking happens by continuous and
simultaneous decision making across various levels.
2.1.2
Decisions
Decisions are concrete occasions for the occurrence of non-knowledge.
“A decision is a choice of action—of what to do or not do.” and “decisions are made to achieve goals,
and they are based on beliefs about what actions will achieve the goals.” (Baron, 2008)
How good a decision was cannot be judged by its outcome (Baron, 2008), but Kahneman (2011)
offers that “most of our judgments and actions are appropriate most of the time” but also warns of
the system 1 influence that can even be observed in the most careful decisions.
Decisions are made all the time, in thinking, private and professional lives by individuals and in
groups. Usually, decisions are embedded into a larger process, but for the sake of science, these
processes have to be reduced to single decision points. In decision science, those decision points are
often not taken from the real world but further reduced by generating results from simplified (game
or game-like) situations.
According to Lawton (2007), decisions, especially when they are political, are inevitably uncertain.
Management, hence also global change management, implies decision making.
Decision Making
Decision making is frequently discussed in literature and various authors offer ways how it can work
in uncertain situations (Barker et al., 2007). With the knowledge explosion, decision making has
become more difficult (Ibisch and Hobson, 2012a). Unknowns and uncertainties are sometimes
perceived as an impediment to effective decision making (Ibisch and Hobson, 2012a).
“To optimally solve a problem, there has to be an optimal solution and a strategy to find it”
(Gigerenzer, 2007a). Decision making is always limited by cognitive or psychological, cultural or
knowledge-related factors (Geiger, Kreft and Ibisch, 2012). One of those limitations are value
systems. Barker et al. (2007) pronounce that individual value systems influence analytical (group)
decisions and that clarifying them is crucial when levels of uncertainty and risk are high. A qualitative
definition of uncertainty along the level of agreement and the amount of evidence (number and
quality of independent sources) is proposed (Barker et al., 2007). Various methods can be used to
reduce complexity in uncertain situations to stay within the cognitive limits of the involved parties
and enable informed and effective dialogues (Barker et al., 2007). Decision-support analysis can help
to deal with uncertainties in cases with a lack of agreement among actors (Barker et al., 2007).
Heuristics show how humans intuitively and routinely deal with unknowns in decision making. Their
systematic study is said to improve decision making. The application of heuristics in suitable
environments is called “ecological rationality”. (Gigerenzer, 2006)
Klein (1999) analysed extraordinary decisions under uncertainty which were taken with expert
intuition – prolonged experience. Klein (1999) asked experienced decision makers, such as fire
fighters, how they decided and why. The role of intuition is then worked out.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
22
Evidence-based decision making promises to increase the use of existing knowledge by systematic
reviews that generate useful evidence (Sutherland et al., 2006). It is problematic in times of global
change and indeterminism: The accessible knowledge explodes (Ibisch, Vega and Herrmann, 2010);
evidence often comes too late (Taleb, 2008); and is often preliminary (Ibisch and Hobson, 2012a);
evidence understands non-knowledge as one-dimensional knowledge gaps (Ibisch and Hobson,
2012a). Evidence cannot deliver what it promises and might eventually cause distrust (Ibisch and
Hobson, 2012a). It attempts to reduce political complexity too much by inviting science as a singular
creator of valid information. There is evidence that evidence gets “disregarded, side-lined or even
discredited if it challenges established practices and vested interests” (Ibisch and Hobson, 2012a),
there must hence also be something like non-evidence and a function in decision processes.
Mistakes always happen and are usually caused by non-knowledge. Mistakes should not impede
decision making: “No space for making mistakes can lead to the behaviour of supporting wrong
decisions for too long instead of correcting the direction and adapting early. It is through allowance
for mistakes that we can understand which option is working and which is not” (Geiger, Kreft and
Ibisch, 2012). Adaptive management encourages decision making based on hypotheses or other
forms of insufficient knowledge. Once mistakes are recognised, it iteratively allows corrective
decisions (Geiger, Kreft and Ibisch, 2012). MARISCO is discussed as an adaptive management tool to
revise, inform and improve decisions and allow transparent decision making: “The unpredictable
nature of managing within complex systems requires vigilance and there is a need for ongoing
evaluation and adaptation throughout the management period” (Ibisch and Hobson, 2014).
A hope for practical advice in decision making is also grounded in econics: “Complications and
frustrations in science and policy are deeply seated in the domain of knowledge and non-knowledge.
If we are able to back cast to first principles of knowledge and non-knowledge in natural complex
systems it may then provide us with more tangible reference points which can then be worked up
into practical tool kits for policy makers and practitioners.” (Ibisch and Hobson, 2012b)
Decision Environment
It is a human trait to search for changes and irregularities in the environment and to assume causality
(as system 1 automatically does) had evolutionary advantages (Kahneman, 2011) and as Taleb (2008)
puts it “we react to a piece of information not on its logical merit, but on the basis of which
framework surrounds it, and how it registers with our social-emotional system. Logical problems
approached one way in the classroom might be treated differently in daily life”. It is a human trait to
anticipate how other people will evaluate our decisions; this social environment therefore matters
(Kahneman, 2011).
In their paper on intuitive expertise, Kahneman and Klein (2009) conclude that intuitive judgement
depends on the predictability of the environment in which the judgement is made and how familiar
the decision-maker is with the regularities of the environment. According to them, e.g. long-term
forecasts of political events are made in zero-validity environments.
Ecological rationality (and the distinct social rationality) is a consequence of how a heuristic matches
with its environment: “we see rationality as defined by decisions and actions that lead to success in
the external world, rather than by internal coherence of knowledge and inferences” (Todd and
Gigerenzer, 1999) or "when generalizing from known to unknown data there is no provably optimal
model” (Martignon and Hoffrage, 1999). Gigerenzer (2007b) offers how to deal with different
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
23
environments and concludes from various studies that imitation can be a good strategy in relatively
stable environments which do often not provide feedback and otherwise lead to dangerous
consequences. Imitation would be ineffective in quickly changing environments. Analysing the
environments, e.g. an institution, in which heuristics are used, led to conclusions on how different
institutions can invite or impede decision making strategies (Gigerenzer, 2006). Organisms are able
to exploit the information structures in the environment by distinguishing if e.g. food is distributed
randomly or with cues (Gigerenzer, 2007a).
Synthesis, Results and Hypotheses
Decisions are processes with several decision points
Decision making usually embraces uncertainties
Evidence-based decision making is limited
Adaptive management allows mistakes and preliminary decisions
The decision environment is crucial for decision analysis
Social reactions are part of the decision environment
2.1.3
Heuristics
George Polya coined the modern use of heuristics with his book ‘How to Solve It’ (1945) where he
distinguished mathematic methods from its content (Baron, 2008). Research into heuristics is carried
out from various perspectives and demonstrates their functioning according to their flaws or their
success depending on the environments. Definitions and understandings of heuristics by the
different authors include:
• “The technical definition of heuristic is a simple procedure that helps find adequate, though
often imperfect, answers to difficult questions.” (Kahneman, 2011),
• “Heuristics are simplified rules of thumb that make things simple and easy to implement. But
their main advantage is that the user knows that they are not perfect, just expedient, and is
therefore less fooled by their powers.” (Taleb, 2012),
• “A heuristic is a process model, that is, a type of strategy rather than a state” (Gigerenzer,
2006) and
• “there is no universal heuristic, but an adaptive toolbox with many building blocks from
which new heuristics can be constructed” (Gigerenzer, 2006).
Heuristics are frequently used mental shortcuts (system 1 reflexive mental operations) supporting
decision makers in complex situations with quick (first) answers. “Heuristics and Biases” is an
approach introduced in the study of human judgement by the psychologists Daniel Kahneman and
Amos Tversky in the 1970s (Tversky and Kahneman, 1974). Tversky and Kahneman discovered with
simple experiments that heuristics can be very valuable but also dangerously flawed by cognitive
bias. A limited number of heuristics are often used in judgement under uncertainty instead of more
formal and extensive algorithmic processing (Gilovich, Griffin and Kahneman, 2002). Kahneman
(2011) argues that he and Tversky investigated on biases because they were interesting and proved
the existence of judgment heuristics. The association of heuristics with biases has often led to the
negative interpretation of heuristics, which is not the scientist´s intention. Gilovich, Griffin and
Kahneman (2002) state that identified heuristics can be debiased, but Kahneman (2011) expresses
that personal control of biases is often not possible.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
24
Gigerenzer (Gigerenzer, 2007b, 1991; Gigerenzer, Todd and The ABC Research Group, 1999;
Gigerenzer and Brighton, 2009; Gigerenzer, 2006) has carried out extensive research on fast and
frugal heuristics that is largely confirmatory of heuristics and simple rules of thumb being better than
complex model-calculations - if applied with ecological rationality. Heuristics are anchored in the
environment and the brain where they tend to be unconscious but can be made conscious
(Gigerenzer, 2007a). Heuristics are specialised for different specific adaptive or inference tasks, e.g.
mate choice, estimation, categorisation or judging a degree of familiarity (Gigerenzer, Todd and The
ABC Research Group, 1999). They are chosen also according to external factors, e.g. time pressure
and success. Gigerenzer and Todd (1999) also promise that in contrast to black-box approaches
"transparent models of fast and frugal heuristics avoid misunderstanding and mystification of the
processes involved.” Therefore, heuristics are tested in real world environments and not compared
to logic or probability. Heuristics can also include inconsistent behaviours, which have advantages if
they are applied by a few individuals (e.g. the rat in the maze, see box).
Figure 5: the quote illustrates "counter-rational" behaviour of rats which is functional in a natural environment.
“Rules can be learned directly, or we can invent them ourselves through a thinking process of
hypothesis testing or reflection. (The use of rules in thinking can be distinguished from the use of
rules to guide behaviour. We may follow a rule through habit without representing it consciously)”
(Baron, 2008). Simple rules are frequently used, personalized and conscious heuristics. They are
investigated e.g. by Eisenhardt (2015a). Those rules can relate to timing, boundaries, etc.; for
instance “If we are not there by 12, we go back”, “only eat what your grandmother would recognize”.
It sets a conscious default option, and most people do not stick too strict with their simple rules but
rather use them as guiding principles. Simple rules are also often observed in companies: “produce
one film every year”. After a simple rule has been decided, it becomes the default option – hence,
“The rat in the maze
A lone, hungry rat runs through what psychologists call a T-maze (…). It can turn either left or right.
If it turns left, it will find food in eight out of ten cases; if it turns right, there will only be food in two
out of ten cases. The amount of food it finds is small, so it runs over and over again through the
maze. Under a variety of experimental conditions, rats turn left most of the time, as one would
expect. But sometimes they turn right, though this is the worse option, puzzling many a researcher.
According to the logical principle called maximizing, the rat should always turn left, because there it
can expect food 80 percent of the time. Sometimes, rats turn left in only about 80 percent of the
cases, and right 20 percent of the time. Their behavior is then called probability matching, because it
reflects the 80/20 percent probabilities. It results, however, in a smaller amount of food; the
expectation is only 68 percent. The rat’s behavior seems irrational. Has evolution miswired the brain
of this poor animal? Or are rats simply stupid?
We can understand the rat’s behavior once we look into its natural environment rather than into its
small brain. Under the natural conditions of foraging, a rat competes with many other rats and
animals for food (…). If all go to the spot that has the most food, each will get only a small share. The
one mutant organism that sometimes chooses the second-best patch would face less competition,
get more food, and so be favored by natural selection. Thus, rats seem to rely on a strategy that
works in a competitive environment but doesn’t fit the experimental situation, in which an individual
is kept in social isolation.” (Gigerenzer, 2007a)
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
25
changing the behaviour (or the rule) would require a new decision. Individuals or groups usually have
a large set of simple rules for different situations (Eisenhardt, 2015a).
Taleb (2012) proposes that heuristics can be an antifragile (see Figure 6) way to do science.
Stochastic tinkering would be such a scientific heuristic. For decision making, those should be what
he calls convex – trial and small error, something that likes variability. Convex issues require small
investments and promise big gains which are a potentially benefiting opportunity. Concave issues
have potentially large losses and a high threat potential.
Figure 6 illustrates the concept of antifragility. The quote is taken from part II of the preface of the book Antifragile
Things That Gain From Disorder (Taleb, 2012).
Examples of Heuristics (and Biases)
A classification and hence “complete” list of heuristics and associated biases is not available (Baron,
2008). Baron (2008) made an attempt and reports of other tries and failures of unifying concepts).
However, some heuristics and biases will be introduced here, as they appeared in the literature and
are reverted to in this thesis.
Heuristic or bias
Descriptive and explanatory quotes
Representativeness
“Judgements influenced by what is typical” (Gigerenzer and Todd, 1999)
Availability
heuristic and
cascade
“Judgements based on what comes easily to mind” (Gigerenzer and Todd,
1999)
“People tend to assess the relative importance of issues by the ease with which
they are retrieved from memory—and this is largely determined by the extent
of coverage in the media.(…)” (Kahneman, 2011)
“An availability cascade is a self-sustaining chain of events, which may start
from media reports of a relatively minor event and lead up to public panic and
large-scale government action.” (Kahneman, 2011)
Anchoring and
adjustment
“Judgments relying on what comes first” (Gigerenzer and Todd, 1999)
“There is a form of anchoring that occurs in a deliberate process of adjustment,
an operation of System 2. And there is anchoring that occurs by a priming
effect, an automatic manifestation of System 1.“ (Kahneman, 2011)
“Some things benefit from shocks; they thrive and grow when exposed to volatility, randomness,
disorder, and stressors and love adventure, risk, and uncertainty. Yet, in spite of the ubiquity of the
phenomenon, there is no word for the exact opposite of fragile. Let us call it antifragile.
Antifragility is beyond resilience or robustness. The resilient resists shocks and stays the same; the
antifragile gets better. This property is behind everything that has changed with time: evolution,
culture, ideas, revolutions, political systems, technological innovation, cultural and economic
success, corporate survival, good recipes (say, chicken soup or steak tartare with a drop of cognac),
the rise of cities, cultures, legal systems, equatorial forests, bacterial resistance … even our own
existence as a species on this planet. And antifragility determines the boundary between what is
living and organic (or complex), say, the human body, and what is inert, say, a physical object like
the stapler on your desk.
The antifragile loves randomness and uncertainty, which also means—crucially—a love of errors,
a certain class of errors. Antifragility has a singular property of allowing us to deal with the
unknown, to do things without understanding them—and do them well. L
et me be more
aggressive: we are largely better at doing than we are at thinking, thanks to antifragility.” (Taleb,
2012)
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
26
“anchors do not have their effects because people believe they are
informative.” (Kahneman, 2011)
Substitution bias
“This is the essence of intuitive heuristics: when faced with a difficult question,
we often answer an easier one instead, usually without noticing the
substitution.” (Kahneman, 2011)
Affect heuristic
“In many domains of life, Slovic said, people form opinions and make choices
that directly express their feelings and their basic tendency to approach or
avoid, often without knowing that they are doing so. The affect heuristic is an
instance of substitution,” (Kahneman, 2011)
"Although analysis is certainly important in some decision-making
circumstances, reliance on affect and emotion is a quicker, easier, and more
efficient way to navigate in a complex, uncertain, and sometimes dangerous
world. Many theorists have given affect a direct and primary role in motivating
behavior." (Slovic et al., 2008)
Halo effect
System 1 exaggerates emotional consistency “The tendency to like (or dislike)
everything about a person—including things you have not observed—is known
as the halo effect.” (Kahneman, 2011)
“The halo effect is also an example of suppressed ambiguity: like the word
bank, the adjective stubborn is ambiguous and will be interpreted in a way that
makes it coherent with the context. (…) Sequence matters, however, because
the halo effect increases the weight of first impressions, sometimes to the
point that subsequent information is mostly wasted.” (Kahneman, 2011)
Recognition
heuristic
Is an example of ignorance based decision making and has a necessity of
ignorance, it uses the less is more effect
(Gigerenzer, Todd and The ABC
Research Group, 1999). For instance: “If you recognize the name of one city
but not that of the other, then infer
that the recognized city has the larger
population.” (Gigerenzer, 2007b)
(Gigerenzer, 2007b)
tested (collective) recognition opposed to more formal
algorithms e.g. on stock markets.
Hindsight bias
"Biased judgments of past events after the outcomes are known (…) tendency
to believe falsely—after the fact—that one would have predicted the outcome
of an event“ (Hoffrage and Hertwig, 1999)
Is common in laypeople and experts and for various types of judgements
(Hoffrage and Hertwig, 1999)
Can lead to overconfidence (Kahneman, 2011)
Consistency bias
“the effect of revising memories in such a way to make sense with respect to
subsequent information.” (Taleb, 2008)
Figure 7 introduces some heuristics and biases which are referred to in this thesis.
According to Gigerenzer (2006) the science of heuristics has three goals: 1. Investigating heuristics
and their building blocks as a toolbox, 2. Ecological rationality – in which environments and
institutions does a heuristic work and 3. Design – the design of environments and heuristics to
improve problem solving. This science has produced various rules and results, e.g. “Cooperate,
forget, imitate” (Gigerenzer, 2007a), “Having more time to search for alternatives increases the
probability of worse decisions” (Gigerenzer, 2007a) or, “in an unknown world, good intuitions have
to ignore information” (Gigerenzer, 2007a). Gigerenzer (2006) proposes that the adaptive toolbox
contains heuristics, building blocks and abilities. The building blocks of strategies are a search rule (in
what order to search cues), a stopping rule (when to stop search) and a decision rule (what to do).
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
27
Synthesis, Results and Hypotheses
Specific heuristics can be turned into simple rules (tools) and used to decide under non-knowledge.
The investigated research on heuristics focusses either on predefined tasks, heuristics or problems
and often takes place in isolated environments.
Tversky and Kahneman (1974) unmasked heuristics by investigating the associated biases with very
simple statistical situations and questions (e.g. a bat and a ball cost 1.10$, the bat costs 1$ more that
the ball. How much is the ball?). Gigerenzer (2007a) for instance tests how a decision is made and
which subliminal heuristics humans or animals automatically use (e.g. to catch a ball) or asks how
good a specific heuristic works compared to more formal, logic or algorithmic processes. ‘Take the
best’ or the ‘recognition heuristic’ are such specific heuristics. Klein´s (1999) approach to decision
making was most similar and partly followed.
These concepts and findings provided methodological ideas and insights but were not useful to the
question and research design at hand. On the one hand, this is due to the specific focus to explore
the unknown in a real-world environment. On the other hand, it depends on the limited capacity to
carry out sophisticated psychological methods.
Simple questions, rules and heuristics can be very effective
Availability heuristic, anchoring, hindsight
Heuristics are frequently used to decide
Decision time, environments and alternatives matter
Environments can be fit to facilitate heuristic decision making
Negative outcomes of decisions are rather attributed to chance and positive outcomes to
ability (the miracle of hindsight)
2.1.4
Knowledge
“Knowledge is a function of the unknown” (Ibisch and Hobson, 2012b)
Cognitive heterogeneity and subsequent constructs can explain much of why knowledge is not
absolute. Cultural and natural knowledge reach into the unknown: we can know about what we do
not know and knowledge manifests. Instances of manifestation are the non-cognitive knowledge
stored in genes or cities (Ibisch and Hobson, 2012a).
Knowledge is specific to individuals or groups and often has personal relevance (Dean, 2008).
Peterson (2008) puts it: “We have some knowledge, but it is limited, changeable, and, above all,
partial. We can hold better and worse knowledge but never any complete, perfect, or final
knowledge. And, whatever we know, it is always from a specific, partial perspective—a standpoint
that fundamentally shapes our way of knowing as well as the content of our knowledge.”
Science is one creator of knowledge and knowledge acquisition is said to be a principal driver of
civilization (Ibisch and Hobson, 2012a). It is acquired through iterative enquiry and tends to cluster
around centres of existing knowledge. This happens with cognitive and non-cognitive knowledge (as
can be observed in nature, e.g. biodiversity hotspots)(Hobson and Ibisch, 2012).
Knowledge has usually a positive connotation (Ibisch and Hobson, 2012a), but some problems can be
found: knowledge is commonly overestimated (Geiger, Kreft and Ibisch, 2012), it can be used
ignorantly or arrogantly (Berry, 2008), it is not equally distributed (Ibisch and Hobson, 2012b), it loses
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
28
validity when the context changes (Nugent, 2008) and much of it cannot be used as it explodes
(Ibisch and Hobson, 2012a).
Berry (2008) offers types of knowledge:
• “sitting-duck knowledge” empirical or provable knowledge of materialists, certainty,
facts, static, smallish once available and acquired it is boring but some is definitely
valuable and useful
• Knowledge as experience and use, which involves uncertainty and risk (do you really
know what the weather is going to be like) and there is no empirical evidence for this
• Traditional knowledge as experience of many people over time
• Religious knowledge, which is only available (and judgeable) to the believers
• Instinct as inborn ‘how to’ knowledge
• Intuitive knowledge (as recognition) – knowing without proof
• Conscience – difference between right and wrong (learned early or innate)
• Inspiration – as a way of knowing that cannot be proved, and imagination as part of this
• Sympathy – intimate knowledge of other people and creatures
• Bodily knowledge – the difference between knowing how and being able
• Counterfeit knowledge or plausible falsehood
Synthesis, Results and Hypotheses
Knowledge can be inventoried
Knowledge is an important driver of civilization but has its limits
Knowledge is particular to individuals, society has cognitive heterogeneity
2.1.5
Non-knowledge
Non-knowledge is multidimensional. It is a lot more than the absence of knowledge; it is larger than
knowledge and grows faster. (Ibisch and Hobson, 2012a)
Figure 8 provides a simple distinction of knowledge and non-knowledge at meta-level.
Known
Unknown
Known
Known knowns
Known unknowns
Unknown
Unknown knowns
Unknown Unknowns
Figure 8 distinguishes knowledge and non-knowledge at meta-level. The first found use of this distinction was by
Rumsfeld (2002) whose speech became popular for the concept.
Non-knowledge is a broad and inclusive term. It embraces other terms such as unknowledge (Taleb,
2012), ignorance (Vitek and Jackson, 2008b) or uncertainty (Gilovich, Griffin and Kahneman, 2002).
Taleb (2012) proposes clusters of unknowns: “(i) uncertainty, (ii) variability, (iii) imperfect,
incomplete knowledge, (iv) chance, (v) chaos, (vi) volatility, (vii) disorder, (viii) entropy, (ix) time, (x)
the unknown, (xi) randomness, (xii) turmoil, (xiii) stressor, (xiv) error, (xv) dispersion of outcomes,
(xvi) unknowledge.”
Ibisch, Geiger and Cybulla (2012) cluster non-knowledge in tangible terms such as knowledge gaps,
blindspots and unknowables. Knowledge gaps are principally knowable and can be recognized or
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
29
unrecognized, blindspots are unrecognized and unknowables are e.g. not determined or lie in the
past. Ibisch and Hobson (2012a) later refine this distinction with the non-knowledge map, diagnose
non-knowledge illiteracy to society and propose non-knowledge literacy as a skill to competently
deal with non-knowledge.
Berry (2008) distinguishes several kinds of ignorance into a non-exhaustive list. Elements of this list
can occur simultaneously. The proposed categories are described in no way as fixed or scientific but
as a reasonable attempt which should at least be the new minimum:
• inherent ignorance caused by the mind – including not only future but also the past (as by
definition we cannot know what we have forgotten)
• ignorance caused by weakness in character, e.g. evidence-based science which ignores useful
knowledge, such as traditions
• Moral ignorance, which is self-induced, often claims objectivity and is related to ignorance
from weakness in character
• ignorance as false confidence, or polymathic ignorance, knowing everything about the past
and how it will stretch into the future
• self-righteous ignorance, failing to know oneself (closely related to polymathic ignorance)
• fearful ignorance (opposing to confident ignorance)
• ignorance from laziness (close to fearful ignorance) – fear of effort and difficulty
• for-profit ignorance – withholding knowledge, as e.g. in marketing
• for-power ignorance – maintained by government secrecy
The concept of risk formalises uncertainties (Barker et al., 2007; Geiger, Kreft and Ibisch, 2012;
Kahneman, 2011). Uncertainty is something unknown that cannot be computed or given a correct
probability and is fundamentally different from calculable risks, where probabilities can be given by
established theories and reliable data (Barker et al., 2007; Taleb, 2008). However, neither in the
literature (Kahneman, 2011; Taleb, 2008; Klein, 1999) nor in the real world are those terms
consistently distinguished. Baron (2008) clarifies that the rules of probability can be applied to
beliefs, even though this is not “correct” calculable probability. Various authors (Kahneman, 2011;
Tversky and Koehler, 2002) point out that measuring uncertainties is always subjective. This
subjectivity and bias depends on the measure which is chosen according to assumptions about
possible outcomes.
Some authors propose distinct causes, levels and understandings of uncertainty. Some get more
explicit than others about the fact that uncertainty is ambiguous, inevitable and expected to
increase. It is frequently discussed where uncertainties come from (change, partial ignorance,
unpredictability and information which is missing, unreliable, ambiguous, contradictory or complex).
It is also discussed that it cannot be dealt with objectively. Dealing with uncertainty requires
accepting the coexistence of incompatible possibilities. Uncertainty often leads to fear, doubt,
delayed decisions (it threatens action) or pursuing of wrong information (Barker et al., 2007; Gilovich,
Griffin and Kahneman, 2002; Ibisch, Vega and Herrmann, 2010; Kahneman, 2011; Klein, 1999; Taleb,
2012).
Randomness is distributed unknowledge. Randomness is practically caused by incomplete
information (Taleb, 2012). Taleb (2008) suggests to accept it and to not assign causality. According to
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
30
Jackson (2008), randomness depends on a limit of perception. This statement is discussed by Jackson
(2008) to be non-scientific and he asks if anyone could make such a statement based on science.
Calling the unknown random would be a first step to its exploration.
Some volatility is the precondition for stability. Volatility and variability are insurances against Black
Swans (Taleb, 2012). Such randomness, including extreme stressors, would at best be decentralised
and is urgently needed for living organisms to prevent them from becoming fragile (Taleb, 2012,
2008). Dean (2008) echoes Friedman´s flat world which decreases local stability, and Taleb (2012)
defines modernity as: “humans’ large-scale domination of the environment, the systematic
smoothing of the world’s jaggedness, and the stifling of volatility and stressors.”
Black swan events, as Taleb (2008) defines them, are high-impact, low-probability events with
retrospective predictability. Examples are a war or the invention of the internet. The problem of
induction, fat tails or the unknown unknowns are similar technical terms. Black swan events “cannot
be dismissed as outliers because, cumulatively, their impact is so dramatic” (Taleb, 2008). Black
swans is supposed to be a concept which maps where the unknown might affect a system. It
illustrates the fragility of knowledge. Taleb (2008) has the understanding that the systems humans
built are so complex that humans cannot understand them. Taleb (2008) underlines hindsight, that
the result of history and often not its generator is seen. Taleb (2008) strongly criticises those
individuals that provide convincing explanations for events that are unexplainable: “we can learn a
lot from data—but not as much as we expect. Sometimes a lot of data can be meaningless; at other
times one single piece of information can be very meaningful. It is true that a thousand days cannot
prove you right, but one day can prove you to be wrong” (Taleb, 2008).
Synthesis, Results and Hypotheses
The literature on non-knowledge opens up a research topic. It draws together knowledge about the
unknown from a wide range of historical and current sources. Perspectives, terms, thought
experiments and classifications were collected.
It appears that the authors of the consulted literature on thinking, decisions and heuristics scratch
the surface of the unknown. This scratching is done with particular terms and perspectives - which
makes it difficult to cluster them. An attempt has been made and its quality probably best represents
the need for this thesis.
Non-knowledge is commonly distinguished by recognition
Volatility, randomness, uncertainty, Black Swan events, ignorance and risks are common
forms of non-knowledge in the used literature
Pattern finding in the unknown is possible
2.1.6
Non-knowledge Literacy
The concept of non-knowledge literacy originated in contrast to knowledge literacy and in light of a
perceived and paradox non-knowledge illiteracy in today´s knowledge society (Ibisch and Hobson,
2012a). Non-knowledge literacy is understood as the ability to competently understand non-
knowledge and use these gained insights as a basis for decisions. Non-knowledge literacy implies
cultural effort and could be understood as a third way opposing to inherent risk-taking strategies or
artificial stability (Ibisch and Hobson, 2012a). An initial understanding would spell: Non-knowledge
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
31
literacy requires an approach that should be broad, systemic and transdisciplinary. It would require
an open debate about the limits to knowledge, including learning to not know, an error-friendliness
and accepting non-knowledge instead of constantly reducing it to closable knowledge gaps. It also
requires learning where non-knowledge is applicable and met with in daily life. As a practical
approach, it should include question based exploration of the (sustainability) topic at hand by making
a systematic knowledge inventory and mapping non-knowledge, including the according relevance
(availability, access, applicability).
Without using the term, the consulted authors have made suggestions how to deal with
uncertainties, non-knowledge or black swans. Quotes on such a handling of the unknown were
collected and are stored in annex VIII.1. Not all techniques would necessarily be called competent
but most claim an intuitive, conscious or skilled handling. The following paragraphs are built on the
quotes in annex VIII.1.
Dealing with uncertainties and other unknowns can be learned and should be taught (Gigerenzer,
2007a; Taleb, 2008). Taleb (2012) reflects that a term (such as non-knowledge literacy) is needed for
a narrative but not for acting (e.g. competently under non-knowledge) and also asks whether “we are
lecturing birds how to fly”.
Non-knowledge should not be an excuse for inaction (Halsnæs et al., 2007; Peterson, 2008; Talbott,
2008; UNFCCC, 1992). Hard-wired, inbuilt or evolved dealing with the unknown is discussed (Geiger,
Kreft and Ibisch, 2012; Kahneman, 2011; Vitek and Jackson, 2008a) and various mind-sets, principles,
decision making systems and heuristics are suggested. Hard-wired and inbuilt dealing with the
unknown is done for example by perceiving more certainty than there is (Kahneman, 2011) –
reductionism is the human condition.
General changes in mind-set are called for, especially that humans should be humble (Peterson,
2008) and learn to live with less certainty (Dean, 2008; Hobson and Ibisch, 2012; Peterson, 2008;
Taleb, 2008). Humans should be prepared (by knowing you cannot be prepared for everything)
(Dean, 2008; Dessai and Wilby, 2011). They should change their attitude towards non-knowledge
(Dessai and Wilby, 2011; Gigerenzer, 2007a; Jackson, 2008; Peterson, 2008; Vitek and Jackson,
2008a). The unknown should be perceived e.g. as a chance, acknowledge it, get hope from it, take
advantage of it or allow curiosity, and a change towards an ignorance-based worldview (Vitek and
Jackson, 2008a).
Kundzewicz et al. (2007) describe that for management, the past cannot simply be extrapolated into
the future. This implies a need for various futures (scenarios) and according uncertainty. Adaptive,
flexible, uncertainty-robust or resilient (not protective) measures are discussed and suggested.
Geiger, Kreft and Ibisch (2012) discuss the need to prepare for possible relevant future scenarios by
hypothesising about all imaginable situations and developments, including emergent properties from
those and suggest tools to generate such scenarios.
An explicit dealing with the unknown by suggesting frameworks (Geiger, Kreft and Ibisch, 2012;
Halsnæs et al., 2007; Ibisch and Hobson, 2014), no or low regret strategies (Dessai and Wilby, 2011)
or calling for robust or resilient, adaptive, flexible, integrative, participatory, proactive, precautious
and preventive strategies (Dessai and Wilby, 2011; Geiger, Kreft and Ibisch, 2012; Halsnæs et al.,
2007; Ibisch and Hobson, 2014; Taleb, 2008) is done by various authors.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
32
Proposed principles, heuristics and rules include:
• Allow uncertainty in politics and keep possibilities open (Peterson, 2008)
• Analyse based on non-knowledge (Ibisch, Vega and Herrmann, 2010)
• Antifragility (Taleb, 2012)
• Approach (non-)knowledge transdisciplinarily (Ibisch and Hobson, 2012a)
• Avoid absolute terms (von Clausewitz, 1832)
• Avoid exposure to small probabilities in incomputable domains (Taleb, 2008)
• Avoid focus in unknown situations (Taleb, 2008)
• Be error-friendly (as cited in Ibisch and Hobson (2012a))
• Blindspotting (Ibisch, Geiger and Cybulla, 2012)
• Challenge your routines in complex situations (Geiger, Kreft and Ibisch, 2012)
• Delay making your hypotheses (Taleb, 2008)
• Distinguish important from unimportant uncertainty (Taleb, 2008)
• Doubt the visibility of immediate causes (Taleb, 2008)
• Have multiple benefits from one choice (Taleb, 2008)
• Heuristics can help in the unknown (Ibisch, Geiger and Cybulla, 2012; Taleb, 2012)
• Know what you (do not) need to know(Gigerenzer, 2007a)
• Maintain redundancy (Ibisch and Hobson, 2012a)
• Maximise options (Dessai and Wilby, 2011)
• Muddle through, let the circumstances guide you in continuous decisions (Giles, 1910)
• Non-knowledge based, broad, systemic, adaptive, scenario-based, precautionary principle,
proactive, resilient, metasystemic, integrative and diverse management (Hobson and Ibisch,
2012)
• Nudge your system 1 out of the important decisions ((Taleb, 2008)
• Observe and monitor constantly (Geiger, Kreft and Ibisch, 2012)
• Omit (Taleb, 2008)
• Precautionary principle (Halsnæs et al., 2007; Hobson and Ibisch, 2010; Klein, 1999; Perry,
2008; UNFCCC, 1992)
• Recognise your unknowns timely (Hobson and Ibisch, 2012)
• Set boundaries (Dean, 2008)
• Take precautions when you have to choose by chance and apply principles of problem-
solving (Klein, 1999)
• Take responsibility (Talbott, 2008)
• Tinker, try undirected (Taleb, 2008)
• Use concepts of econics (Ibisch and Hobson, 2012a)
• Use your intuition even though it ignores information (Gigerenzer, 2007a)
Rather protective measures include: limit interactions with the unknown (Dean, 2008), act decisively
(Klein, 1999) and use non-knowledge to maintain your position (Giles, 1910). Some bold ideas by
Taleb (2012, 2008) suggest stabilizing the system by injecting unknowns, allowing to break what
needs to be broken and to make many small mistakes.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
33
2.2 The Post-normal Science Diagram
Reviewing the literature on decision making, it became clear that any decision depends on its
environment, especially on the (un)certainty it provides. The post-normal science diagram (Figure 9)
provides a heuristic assessment of the environment. “When a problem is recognised as post-normal,
even the routine research exercises take on a new character. For the value-loadings and
uncertainties are no longer managed automatically or unselfconsciously” (Funtowicz and Ravetz,
2003).
Post-normal science is made for science-related policy and promises to retain complexity. “Contrary
to the impression that the textbooks convey, in practice most problems have more than one
plausible answer, and many have no well-defined scientific answer at all” (Funtowicz and Ravetz,
2003). The Post-Normal Science diagram focuses on aspects of problem solving for issue-driven
science where “typically facts are uncertain, values in dispute, stakes high, and decisions urgent”
(Funtowicz and Ravetz, 2003). It is designed for situations where “we must make hard policy
decisions where our only scientific inputs are irremediably soft” (Funtowicz and Ravetz, 2003).
Figure 9 The post-normal science diagram (Funtowicz and Ravetz, 2003) provides a framework to map systems
uncertainty against decision stakes. It postulates that applied science can be used when both are low, professional
consultancy when at least one is medium and post-normal science when systems uncertainty or decision stakes are high.
The post-normal science diagram is one concept used in the questionnaire.
The post-normal science diagram can assess the environment uncertainty
A third dimension might be missing in the post-normal science diagram
2.3 Exploring the Non-Knowledge Map for Decisions
The original non-knowledge map (Figure 3) was explored by tinkering. It was used in the trial run
questionnaire and for additional decisions from the author’s life. This application and use was
necessary to understand the map. A need to adapt the map for decision situations was discovered.
This was basically created by what Baron (2008) describes and discusses as reflection: “Reflection
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
34
includes the essential work of philosophers, linguists, mathematicians, and others who try to arrive
at general principles or rules on the basis of evidence gathered largely from their own memories
rather than from the outside world. (…) In reflection, the search for evidence is more under the
control of the thinker than in diagnosis and experimental science; in particular, thinkers can direct
their memories to provide evidence either for or against a given possibility (in this case, a
generalization).”
One line of enquiry was inspired by antifragility and not robustness being the opposite of fragility
(Taleb, 2012). The gradient of threat potential is expressed as reaching from a high threat potential
to potentially benefiting (opportunity). It was tried to attach antifragile values for other dimensions.
The non-knowledge map (Figure 3) needed some further modifications for being applied in decision
situations. For instance, in the case of “knowability” it is not clear if something is knowable for the
decision maker in the decision situation or generally knowable.
It was found that the non-knowledge map can also be displayed in a table format (Figure 10), which
facilitates the use in the questionnaire and its evaluation. For evaluation, either ends, both poles, or
the gradients of each dimension, have been assigned -2 and 2. The numbers only express that the
two ends of each dimension are bipolar; it does not express positive or negative value.
Figure 10 shows the simple table format non-knowledge map. For evaluation numbers from -2 to 2 have been assigned.
The numbers only express that the two ends of each dimension are bipolar.
The adapted table format non-knowledge map (Figure 11) has some parts of the gradient more
closely defined (displayed as 1 and -1). The order of dimensions has been changed during the course
of the interviews and “social distribution” and “geography” have been removed from the map, as
they do not apply in decision situations. ‘Solution relevance’ was renamed into (decision) ‘relevance’.
Those seven dimensions with the respective gradients are explained in conjunction with the
interview results in 3.3.1.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
35
-2
-1
0
1
2
Temporality past-related
extends through
present
future-related
Ambiguity
clear
(unambiguous)
retrospectively
ambiguous
ambiguous
Knowability knowable
knowable with
investment
now unknowable unknowable
Relevance
not relevant
would be relevant
highly relevant
Threat
Potential
potentially
benefiting
high threat
potential
Intentionality unintended
neglected
(non)knowledge
intentional
Recognition
fully
recognized
retrospective
recognition
unrecognized
Figure 11 shows the adapted table format non-knowledge map as used in questionnaire version 1.4.
2.4 Results and Hypotheses from Theoretical Part
The following results and hypotheses are to be principally investigated in the empirical part.
A list of potential elements for non-knowledge literacy (2.1.6)
Heuristics are functional in conjunction with their environment (2.1.3)
An adapted non-knowledge map that can be applied to decision unknowns (2.3)
An adapted version of the post-normal science diagram can capture the environment
(2.2)
The following questions and points are secondary aspects of investigation in the empirical part. They
stem from the syntheses of the preceding chapters.
(How) is system 1 thinking and system 2 thinking used in work decisions?
Are effective simple questions, rules and heuristics used in work decisions?
Decisions are processes with several decision points
Decision making usually embraces uncertainties
Evidence-based decision making is limited
Adaptive management allows mistakes and preliminary decisions
The decision environment is crucial for decision analysis
Social reactions are part of the decision environment
Availability heuristic, anchoring, hindsight
Decision time, environments and alternatives matter
Environments can be fit to facilitate heuristic decision making
Negative outcomes of decisions are rather attributed to chance and positive outcomes to
ability (the miracle of hindsight)
Knowledge can be inventoried
Knowledge is an important driver of civilization but has its limits
Knowledge is particular to individuals, society has cognitive heterogeneity
Pattern finding in the unknown is possible
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
36
3. Conservation Phase: Empirical Part
Results and hypotheses generated in the theoretical part were conceptualised into a questionnaire.
This questionnaire was used in 21 interviews. It generated variables for 39 decisions and 40
unknowns. Those variables were statistically analysed. Qualitative analysis of separate interviews,
decisions and unknowns was also done. The interviewees are presented first.
3.1 Interviewees
A Global Change Manager is a person who has successfully completed the International Master Study
Programme “Global Change Management (MSc)” at Eberswalde University for Sustainable
Development (before 2010 the University of Applied Sciences, Eberswalde). The GCM study
programme started in September 2006 out of the need to translate knowledge generated about
global environmental changes into applied concepts. Detailed information about the rationale,
students and the curriculum are given in annex VII.2.
It is assumed that Global Change Managers are open-minded and comparably easy to find and
approach. Their work decisions are also comprehensible for the author. More experienced decision
makers might also be more difficult to talk to for the author, as they would know a lot more (either
about their discipline in which they are so grounded that it is hard to get them out or about non-
knowledge itself).
21 interviews were carried out. These 21 interviews produced analysis of 39 distinct decision
situations and 40 non-knowledge maps. In three interviews, due to time constraints or lack of
decisions, only one decision was analysed. Hence there are 21 first decisions and 18 second
decisions. In three cases a different aspect of the non-knowledge was mapped instead (2) or
additionally (1). One interviewee could not name anything unknown in a decision situation so that for
one decision no non-knowledge was mapped.
The interviews took place on 15 days within three weeks from August 20 to September 8, 2015. The
author had a maximum of three, sometimes two but usually one interview per day. One interview
was split between the two decisions and took place on two different days. Three interviews were
done by phone and two with Skype video calls. All others were Skype calls. The recording failed in
two cases.
The interviews took on average 45 minutes. The shortest and the longest interview treated only one
decision and took 20 and 95 minutes respectively. The other interviews took between 30 and 60
minutes. Usually interviewee and author talked before and after the interview so that the calls took
longer than the interview. The author knew (e.g. from lectures, student jobs or internships) ten of
the interviewees before the interview and was familiar with two decision situations and four decision
environments.
12 interviewees were female and nine were male. Interviewees were aged between 28 and 47 with
an average of 34 years. They are nationals of seven different countries, 15 European (including 14
Germans) and two North American, South American and Asian nationalities each.
Interviewees began studying GCM between 2006 and 2012, with 2-4 alumni from each year. They
finished between 2009 and a prospective end of study in 2016. One interviewee studied but never
finished GCM. For the data analysis it was assumed that this person finished GCM after the average
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
37
study duration. The average study duration was 3 years (12 interviewees); five were below average
and three above average. At the beginning their calculated age was between 24 and 44, with an
average of 28 years. Before studying GCM the interviewed alumni held 13 different degrees,
including two with two degrees. The majority (16) held a degree in various environmental sciences;
half of those in International Forest Ecosystem Management (BSc) from Eberswalde University for
Sustainable Development. Interviewees also held a degree in various social sciences, engineering or
economics. Before studying GCM they had an average of 3.5 years of work experience with 16 below
average, including the minimum. The minimum was six interviewees that did not have any work
experience. Five were above average; two of those having the maximum of 12 years work
experience.
Four to seven interviewees currently work in research institutions, NGOs, companies or
administration, respectively. Their positions (in order of decreasing frequency) are called manager,
researcher, officer, freelancer, advisor and assistant. However, these are not representative
categories, as terms are not used consistently, e.g. a manager can be a project manager or a regional
manager. They have held this position for an average of three years with the minimum of less than
half a year (2) and the maximum of six years (2).
The average total calculated work experience is 6.5 years with one person each holding the minimum
of 1.5 years and maximum of 16 years of work experience.
3.2 Methods
A questionnaire was conceptualised for the trial run and subsequently adapted. Results for the trial
run can be found in annex VII.3. 20 interviews were carried out with the questionnaire. The results
were analysed quantitatively and qualitatively.
3.2.1
Questionnaire
The questionnaire versions can be found in annex VIII.2. Building blocks of the questionnaire are:
(1) General information
(2) Evidence question
(3) Decision-related questions
(4) Mapped non-knowledge
(5) Post-normal science diagram inspired questions
(6) Evaluative closing questions
The questionnaire is built to analyse two decisions. This produces two sets of answers for all building
blocks apart from ‘general information’ and ‘evidence question’. The building blocks and questions
are not numbered in the following. It is hence suggested to keep the above bullets in mind, as they
are used throughout the thesis. The same is suggested for questions within the building blocks.
The questionnaire was built around the non-knowledge map. The literature on decision making made
clear that any decision depends on its environment, especially on the (un)certainty it provides. The
post-normal science diagram (Figure 9) provides a heuristic assessment of the environment.
The questions around the non-knowledge map and the post-normal science diagram are based on
common sense knowledge, experiences, literature and experiments. Generally categories and
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
38
questions remain non-exhaustive. The study of heuristics shows that simple questions, rules and
heuristics can be very effective (Gigerenzer, Todd and The ABC Research Group, 1999). Hence the
questionnaire was guided by simplicity, and few categories, questions and options were offered.
Throughout this thesis, “I” indicates potential or real thoughts and words of an individual. If these
thoughts are not in quotation marks (e.g. I wanted it), they refer to potential thoughts and often
comprise fixed categories for analysis. If it is in quotation marks (e.g. “I did not know if this was good
or bad”) it is a real sentence and wording used by an interviewee. If it is made clear that an
exemplary interview is stated, there are not quotation marks as the I refers to an edited version of
the wording used by the interviewee. There might be a few deviations from those rules but they
should become clear from the context. The author is never referred to as “I”.
General information
As it is common practice, personal characteristics such as name, age, gender and subject-specific
information such as duration of GCM studies, employer, position and former studies were asked of
the interviewee. Work experience was understood as full-time paid employment.
Evidence question
An initial thought provoking question was asked about evidence. It should answer the question how
receptive alumni might be to non-knowledge and “if they have all drifted into evidence-based
decision making” (Ibisch, 2015).
Decision-related questions
As decision retrieval from memory proved difficult in the trial interviews, interviewees were asked
before the interview to think of two global change related decisions they had taken at work. Both
decisions should be real and concrete decision taken by the interviewee (alone or in a team) at work.
As many interviewees considered their decisions to be “too small” and “not actually global change
related”, any other decision situation from work was accepted. Those two decision situations
retrieved from memory are probably influenced by the availability heuristic (as described e.g. in
(Kahneman, 2011).
“In reality there are ill-defined goals” (Klein, 1999) and a decision is usually a process consisting of
many decision points. Some questions relate to the decision process, others to a concrete decision
point. Hence it was necessary to have interviewees define a concrete decision point. The author
rephrased a concrete decision point she understood from the interviewee’s explanations (sometimes
lengthy monologs about a complex process), e.g. “so, your decision was to decide if to do x” or “your
decision was to suggest x to y”. If the interviewee did not agree with the suggested decision point,
the procedure was iterated until a concrete point was reached.
For each decision situation, the interviewee was asked to describe how the decision was taken
(process), why it had to be taken (point), if alternatives could be influenced (point) and how long the
decision process took. Furthermore, the interviewees were asked to name some of their knowledge
and non-knowledge at the specific decision point.
Asking how the decision was taken aims at distinguishing between largely system 1 and largely
system 2 informed decisions and was initially conceptualised as “the heuristics question”. Research
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
39
about heuristics usually focuses on predefined situations or predefined problems (2.1.3). This
prescriptiveness was to be prevented in this thesis as it is understood to anchor interviewees. Hence,
the possibility remained to ask an open ended question with the hope to later find patterns in these.
These patterns were clustered later. When interviewees struggled with the question, they were
offered to answer if it was a rather intuitive or a rather rational decision.
How decisions are taken is often unpredictable (Ibisch and Hobson, 2012a; Taleb, 2008), this could be
attributed to human rationality and irrationality (Ibisch and Hobson, 2012a).
Several “reasons” for how the decision was taken were derived from the trial interviews and the
decisions used during tinkering. At the beginning they were constituted of: it was clear (strong
system 1 influence), I wanted it, to satisfy someone else (guided by circumstances), rational analysis,
unhappy about alternatives (guided by circumstances), it meant progress (guided by higher goal), it
was part of a package (guided by higher goal), not to be blamed later(guided by circumstances),
precautionary principle (guided by higher goal), muddle-through (not visibly guided), “I always do it
like this” (guided by routine), “I was taught to decide like this” (prescriptively guided by rules). These
suggestions remained in the questionnaire during all versions but were usually not read out, as the
list was too long and the options did not quite apply for the named decisions, which were from a
working environment.
Decision time is an important factor, as many decisions that are taken quickly use simple heuristics
(Rieskamp and Hoffrage, 1999) or are system 1 inspired (Kahneman, 2011). Gigerenzer (2007a) found
that “having more time to search for alternatives increases the probability of worse decisions.” The
time required for a decision is distinguished as: I took the decision right away, I took some finite time
(usually defined by the situation) or I took all the time I needed.
Many interviewees would fit their answers of why they had to take the decision into the categories
of it was a requirement, it was my job to decide this, I had to decide because I had the idea to do this
or as a response to an unfolding situation. If interviewees struggled with the question, they were
asked to answer if it was for example a requirement, a response or an idea (according to the author´s
understanding gained about the situation so far). Hereby it became clear, and was usually rephrased
and asked separately, if the interviewee could influence the alternatives.
Interviewees were then asked what they knew in the concrete decision situation (point) to get them
into thinking in terms of knowledge and non-knowledge. It was not meant to be a full knowledge
inventory and not much focus was laid on this question, the author rather pushed the interviewee to
go over this question quickly. Often the author offered knowledge from information given by the
interviewee beforehand and was satisfied with one or two additional “pieces of knowledge”
(knowns). Usually interviewees would rephrase what they had said before, but specifically focused
on the decision point.
Interviewees were then asked what they had not known in the decision situation. Here they were
pushed to name as many unknown aspects as possible, or relevant. Sometimes the author offered
non-knowledge from information given by the interviewee when answering previous questions to
speed up the process. After all recognized and available non-knowledge was named, the author
either asked the interviewee which unknown aspect would be considered most relevant and should
be investigated more in depth, or occasionally defined the aspect for further investigation herself.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
40
This was in order to map a larger variety of unknowns. This aspect was then taken on to the non-
knowledge map.
During the first interviews it became clear that it matters if a decision is final or if it can be adapted if
needed. Hence it was noted in the spreadsheet whether the decision was final or preliminary. It was
also noted if the decision was taken alone, supported (a single decision maker has consulted the
problem with a second party) or in a group.
It has to be kept in mind that answers were subjective and retrospectively distorted - biased by
hindsight (2.1.3). However, the interviewee was constantly reminded to answer questions as it was
perceived “in the decision situation” or “when you took the decision”.
Mapped non-knowledge
The decided piece of non-knowledge was then set in relation to the dimension’s gradients, e.g. “was
it intentional or unintentional that you did not know if…”. See Figure 11 for the dimension´s
gradients. Questions were formulated and asked in the order of dimensions on the printed map.
Formulation of questions was sometimes not easy as they used the double negative. Interviewees
usually had to be reminded of the unknown talked about in every question. An example formulation
is “That you did not know if this was good or bad, was it intentional or unintentional?”, the
´neglected´ gradient was only asked when it seemed to apply.
In the category “type of non-knowledge” the author had to write down a classification for the non-
knowledge that was analysed. It was supposed to train the non-knowledge classification abilities of
the author. It was also designed to rephrase the non-knowledge just described by the interviewee
and this should help the interviewee get into thinking about non-knowledge (which was quite
challenging for most interviewees). However, the types of non-knowledge were not convincing so
that this category was not read to the interviewee. The initial list of potential types was: about (local)
facts, a dynamic situation, another person´s behaviour, personal preferences, probabilistic,
something undetermined, alternatives, random or by chance … . This list was to be changed in the
course of the interviews but this was not successful as the intermediate categories future, others,
information and value were disproven.
Post-normal science diagram inspired questions
The post-normal science diagram (2.2) was used as a base to formulate questions according to the
systems uncertainty and the decision stakes. Systems uncertainty was renamed as environment
volatility. During the process of interview design, in personal communication with the supervisors, it
became clear that the idiosyncrasy, so the characteristics that belong to this person could be a
possible third dimension. Initial ideas were very close to heuristics, to identify if the person had
decided rather rational, more intuitive or according to some rules or routines. During discussion with
the supervisors this turned into “condition/fitness” and finally “idiosyncrasy”.
As a direct placement on the graph did not prove very useful in the trial interviews, three questions
with three choices for each dimension were included in the questionnaire. The values 1, 2 and 3
(low/small, medium, high/large) were attached to the three choices. Then the average was
calculated for the three dimensions of decision stakes, environment volatility and idiosyncrasy. Those
are discussed in the following paragraphs.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
41
Environment volatility is defined by the topical, institutional and personal environment. The topical
environment refers to the topic of the decision. The institutional volatility refers to how established,
large and structured an organisation is, it is assumed that e.g. ministries, city administrations or large
companies provide a certain decision structure that has evolved or functioned over time (although
these are more prone to be hit by negative Black Swans). In contrast to this, freelancers or small
NGOs with flat or no hierarchies, few years of experience and little structure are assumed to be more
uncertain or volatile decision environments. The assessment of uncertainty of decision situations
shall also give initial ideas about how institutional environments could be adapted (Gigerenzer,
2006). The personal environment asks about the relations at work. If a relation to a boss or
colleagues is tense or the boss´ or colleagues´ reactions are very unforeseeable, it might influence
the freedom with which decisions are taken: “Many of us spontaneously anticipate how friends and
colleagues will evaluate our choices; the quality and content of these anticipated judgments
therefore matters” (Kahneman, 2011).
In the interviews, the decision stakes were defined by impact on people, investment and impact over
time. Impact on people relates to the influence the decision has on other people; was it a decision
which influenced only the decision maker or did it have a global reach? It sets the context from an
anthropocentric perspective. Indirectly, it incorporates impacts on the natural environment, as these
would feed back into the system and impact people, e.g. on a global or a local scale. The investment
variable refers to the decision stakes relating to time and monetary or other material investments
implied by the decision. Those are usually considered scarce resources and are often used as
indicators or excuses. Investment in terms of ideas or thoughts is not always abundant but difficult to
be measured. It is also covered in the idiosyncrasy questions. Impact over time asks how long the
impact will be observable over time; a few days would lower the decision stakes whereas a decision
that will have an impact over many years is considered to have higher decision stakes. However, the
investment is always dependent on the institution that is implementing the decision. It can be very
different for a small project of an NGO to a global endeavour. This cannot be captured.
The decision maker´s idiosyncrasy is assessed by questions on mind-set, values and non-knowledge
awareness. Assessing this objectively is not possible (in this thesis). However, to capture the
idiosyncrasies, the interviewee is asked for the influence which his or her values and mind-set had on
the situation. Additionally the author noted the perceived non-knowledge awareness of the
interviewee. This gave a first impression on how the interviewee deals with unknowns. Some
interviewees instantly agreed to talk about unknowns, gave examples and answered questions
quickly and competently. Others had difficulties in understanding the questions concerning the
mapping of concrete unknowns. The highest score was given when the use of words and phrases
demonstrated non-knowledge awareness.
Evaluative closing questions
Closing questions ask how happy, worried or non-knowledge aware the interviewee was when
deciding, how sure they were at the interview day that their decision was good or bad and if they
think this was based on their ability to make good decisions or if it was by chance. These provocative
questions are derived from the heuristics and heuristics and bias literature, where in posterior
evaluation, events with negative outcomes are rather attribute to chance and positive events to
ability. By the use of hindsight, the mind can deal with it.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
42
3.2.2
Interviews
The core material for this thesis was generated in 20 interviews with Global Change Managers. For
information about the study programme see annex VII.2. They were interviewed for 30 minutes via
Skype. Interviews were audio recorded. A questionnaire was filled out by the author and the data
later transferred into a spreadsheet to document the results. The author took additional notes about
the interviewee´s comments and impression of the interview in a separate document. This document
served as a further source of information.
20 interviews were heuristically set. It is a manageable amount that might provide a minimal level of
statistical representability. It can also be assumed that in human environments, information
structures are observable rather quick. Kahneman (2011), for instance, reflects that biases he
observed in his own thinking proved usually correct in further testing when his companion Tversky
fell into the same trap. Hence, it is not attempted to give statistically or empirically relevant
unknowns or heuristics from 20 interviews but if some patterns would emerge in 20 interviews (40
decision situations), these might be worth more in depth investigation.
The objective of the interviews is to apply the non-knowledge map to the non-knowledge of
different people in different situations and see if some variables allow clustering. It is not the aim to
compare between interviewees, make statistical inferences or present a complete picture. It is
merely an extension of the reflections and tinkering with the map, an extended peer community that
includes people with different views, experiences and decision situations. The complete
questionnaire is presented in the following but the core information used in this thesis is the non-
knowledge map.
Scheduling of Interviews
In a joint effort of the author, her supervisor and the coordinator of the study programme, the
names of GCM alumni they knew were retrieved from memory. A non-exhaustive list of names from
every GCM generation was then checked for familiar names. One interviewee suggested other
alumni that were then contacted. In this way, 36 GCM alumni were personally contacted via email,
mainly by the author´s supervisor.
Scheduling of interviews was easy and replies quick and sometimes positive “I would be happy to
take part and am glad to hear that you are working on such an interesting topic, which is quite
relevant to my work – even though we don’t discuss these issues directly (they are sort of just there,
like the elephant in the room).” Some interviewees asked for a short introduction of the research
topic and the thesis´ aims, which was then quickly summarised. After the interview, 13 interviewees
expressed interest in the results.
Recording of Interviews
All interviews were audio-recorded as backup. The records were used to clarify certain information
and re-listen to difficult episodes. They were transcribed and systematically analysed or reheard in
case of doubt and for discursive interviews.
Development of the Questionnaire
The questionnaire was developed from an initial trial run version 0.0 to 1.4. It always contained three
parts: basic information about the interviewee, a non-knowledge section about two concrete
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
43
decisions and corresponding maps as well as questions about the decision and its environment along
the lines of the adapted post-normal science diagram. The steps of questionnaire development are
displayed in annex VIII.10.
Use of the Questionnaire
The questionnaire was used as guidance for the author and to note down answers. Especially in the
earlier versions (up to 1.0), the author often occasionally deviated from order in asking questions and
formulated questions more freely than they were written down. This allowed improving questions
and change order to increase understanding. A guided change in wording and deviations from
written down questions happened, e.g. in cases where the interviewee had difficulties in
understanding questions and abstract thinking. When an option was obvious from previous
information given by the interviewee, often just this option was offered. However, in most cases the
questions and options were read to the interviewee.
Decision 1 was noted on the left side of the sheet with blue colour and the second decision in black
colour on the right side of the sheet. Solid lines and circles were given for the first decision and
dashed lines and rectangles for the second.
3.2.3
Qualitative Analysis
Qualitative research in form of case studies allows generating new theories (Eisenhardt, 1989).
Disciplines such as ethnology, pedagogy or sociology use qualitative research concepts (Breuer,
2010). In today´s psychology they are rare, although regional differences exist (Breuer, 2010).
Exemplary interviews, decisions and unknowns are presented and answers to the open questions
clustered and coded.
3.2.3.1
Exemplary Interviews, Decisions and Unknowns
For the qualitative analysis, each interesting decision was marked in the spreadsheet and interesting
or problematic dialogues that evolved during the interview were noted. Results (anonymised and
alienated) are displayed in the following sections:
• 3.3.2: one decision with all answers.
• 3.3.6: six selected decisions with their characteristics. The decision point, the characteristic
and associated non-knowledge are shortly described in the I-form.
• 3.3.6: content from five interviews with discursive dialogues about non-knowledge and the
asked questions. Many direct quotations are presented, as those interviews were transcribed
to extract the information generated.
• 3.3.1: unknowns (mainly from the interviews) illustrate the dimensions and are presented in
conjunction with their decision point.
Direct quotations from the interviewees are presented in inverted commas without attributing it to
the speaker to maintain anonymity.
All information from the interviews was recorded in a spreadsheet. This was the data pool for the
following analysis. All qualitative information was clustered if necessary and transferred into
numbers to allow quantitative statistical analysis with the principal component analysis (PCA), bar
charts and linear bivariate models.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
44
3.2.3.2
Coding of How did you Decide?
The “heuristics question” was analysed in depth. For that purpose, information explicitly given by the
interviewee was used to identify keywords. Only those keywords were taken on and their implicit
influence on the decision was assessed. The frequency of keywords was counted. All keywords were
then clustered and linked with literature. The value of (sub-)clusters for each unknown was then
calculated and transferred into the spreadsheet.
This procedure is similar to an interpretation of the psychological grounded theory methodology:
"for some scholars, grounded theory building simply means creating theory by observing patterns
within systematically collected empirical data.” (Eisenhardt and Graebner, 2007)
Explicit Keywords
The text in the spreadsheet (e.g. gut feeling and investigations until evidence; rational, plus a bit of
intuition; rational analysis and discussion) was transferred into a new spreadsheet and analysed. For
every new piece of information (keyword), a new column was created and ticked. If keywords
reappeared, they were ticked in the existing column. Some minimal clustering was done if the
wording was different but the meaning the same (e.g. the combination of sense and plausibility in
one column). An understanding of the terms is given in annex VIII.2.
Examples are: “I decided intuitively”, “I thought about it”, “I analysed it”, “I had a gut feeling”
Implicit Keywords
The keywords generated from explicit information were ticked for all implicitly given information by
the interviewee for the specific unknown (rows). This information was given in answers to other
questions or could be read between the lines. In case of doubt the recording was consulted. If doubt
remained, a precautionary approach was taken and the row was not ticked.
Frequency of Keywords
It was counted how often implicit and explicit information was ticked for each decision (column) and
keyword (row).
Clustering of Information
All keywords were transferred into a separate document (compare annex VIII.2) and clustered in
several steps until the author (and her supervisor) was satisfied and the information could be
sensibly linked to the literature. This clustering was applied to the spreadsheet, so that the
spreadsheet contained different (sub-)clusters and categories for analysis with the according values
for each decision.
Examples are: largely system 1 informed (cluster), largely informed by accessible information (sub-
cluster), informed by experience (sub-cluster)
Additional Calculation
It was calculated for each (sub-)cluster if and how many of the corresponding keywords were ticked
so that the clusters could be analysed in the PCA. It was distinguished how many and if any keywords
were ticked for each (sub-)cluster (compare annex VIII.6).
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
45
3.2.3.3
Coding of Decision Conditions
No distinction between explicit and implicit information was made, as useful categories became clear
quite quickly. Some basic clustering was carried out. For the category “it was my job to decide this” it
was assigned if it was a part, an important part, the core or not “my” job and each category was
given a numerical value. The other options were ticked (1 if yes, 0 if no).
Examples are: it was a requirement, an idea, requested, a directive
3.2.3.4
Coding of Knowledge and (Mapped) Non-knowledge
The text in the spreadsheet was transferred into separate documents for knowledge, non-knowledge
and mapped non-knowledge and clustered independently in various steps (a combined document of
the final results can be found in annex VIII.8). In a last and combining step, they were clustered into
just four categories (VIII.6):
• circumstances (e.g. resources, time, capacities),
• consequences (what will happen if I do this or the alternative),
• complementary knowledge (I was in a similar situation before and can apply this knowledge)
• comments about the importance of the known aspects and respectively the extent of the
non-knowledge.
In case of doubt, the recordings were consulted. For each decision it was counted how many
(un)knowns for each of the four clusters were explicitly given by the interviewee.
Complementary (non-)knowledge can be grasped from this quote: “We may transform what we get
in a variety of ways to make it applicable to our situation. This is the important mechanism of
analogy” (Baron, 2008).
Those are, however, no strong indicators as only the explicit (un)knowns were analysed. Hence, the
mapped non-knowledge was additionally divided into four more functional categories.
3.2.4
Quantitative Analysis
Bar charts were used to quantitatively analyse the distribution of unknowns across dimensions. A
linear model analysed the distribution of the decisions on the three-dimensional post-normal science
diagram and a principal component analysis was carried out by iteratively jack-knifing the initial data
input set with 139 variables for each decision.
The programme PAST (Hammer, Harper and Ryan, 2001) was used to do a multivariate principle
component analysis (PCA) of different data sets from the spreadsheet. A PCA finds components that
account for data variance. Components are hypothetical variables built from linear combinations of
input variables. Results are displayed by plotting two components (usually the first two components).
Those can provide hypotheses based on objective data analysis and might provide important
components to correlate with other variables (Hammer et al., 2001). The “correlation” function was
used to streamline data which were in different units. In some data sets groups were specified.
Results from the tabs “summary”, “scatter plot” and “scree plot” were generated and pasted into a
separate document (compare VIII.5). The initial data set was jack-knifed in roughly ten steps. If the
scatter plot produced visible hypothesis, groups were applied to it.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
46
3.2.5
(Non-)availability of Data
The annex (VIII.6) contains the raw data (numerical values) from each interview. Recordings and filled
questionnaires contain sensitive data and are not published.
3.3 Results
Qualitative evaluation of the interview results was carried out. Results are presented in relation to
the main hypotheses and results from the theoretical part:
• Collected unknowns and the non-knowledge map for decisions
• Answers to the heuristics question “how did you decide?”
• Use of the listed potential elements for non-knowledge literacy
• The post-normal science diagram
Subsequently, all answers for one exemplary interview are displayed, some exemplary decisions are
depicted and contents from five discursive interviews are presented. Answers to the remainder of
the questionnaire are presented last.
For the quantitative analysis, all collected information in the spreadsheet was coded and statistically
analysed. Furthermore the decisions were placed on the post-normal science diagram and the
distribution of unknowns on the non-knowledge map was presented as a bar chart.
3.3.1
Collected Unknowns and the Non-knowledge Map for Decisions
The unknowns collected in the interviews and the selected unknowns to be mapped are presented.
The distribution of mapped unknowns among the dimensions is shown as well as reflections about
each dimension.
Inventoried Non-knowledge: What did you not know when you took the decision?
Interviewees phrased their unknowns e.g. as “I did not know” or “I did not have much insight in”.
Some unknowns were: the changes of what will happen in the forest, the legal architecture, what is
financially feasible, the intention of some actors, if they would like it or where to get the information.
To maintain anonymity of the interviewees, not more context than this is given for the unknowns.
Usually, interviewees could give their unknowns as quick as their knowns, only in two cases
interviewees said that there had not been anything which they did not know in the decision
situation. In one situation the interviewee could be convinced and found several (decision irrelevant
but closely related) unknowns. In the other of the two cases, there was no unknown given by the
interviewee.
In 30 decision situations interviewees mentioned at least one unknown about the circumstances. In
20 and 13 cases respectively, they told what they had not known about the consequences and their
lack of complementary knowledge. Six commented on how much they did not know (e.g. “many
uncertainties”, “there was nothing relevant I did not know”, “many blindspots”). Circumstance non-
knowledge was also illustrated most, interviewees gave a total of 76 unknowns, and one interviewee
gave five aspects of their circumstance non-knowledge but most gave zero to four. In contrast to this
33 and 23 aspects of consequence and complementary non-knowledge were given, usually zero to
three for each decision. In only six cases interviewees gave an estimate of how much they had not
known. 26 did not give any complementary non-knowledge, 19 gave no consequence non-knowledge
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
47
and nine gave no circumstance non-knowledge. For the average decision 3.8 unknown aspects were
given.
Mapped Non-knowledge
Unknown circumstances were mapped in 24 cases. 13 cases mapped unknown consequences. In
three cases, a lack of complementary knowledge was mapped. Four of those were value unknowns
(e.g. “I did not know if it was good or bad”), 15 each were non-knowledge about others (“I did not
know someone´s intention or reaction”) and about the future (a dynamic situation) and six were
about unknown information (it can be information that will only be generated in the future).
The maps are usually different but there are two pairs, one triplet and twice four times the same
map. One pair was generated by one interviewee for two unknowns belonging to the same decision
situation but all others are from different interviewees. The only difference between the two sets of
four similar maps is the threat potential.
Pair (2.20, 7.20): future related, clear, now unknowable, would be relevant, high threat potential,
unintentional and recognised.
Pair (12.11 and 12.22): Extends through the present, clear, now unknowable, would be relevant,
potentially benefiting, unintentional and recognized
Triplet (5.20, 21.20, 9.10): Extends through the present, ambiguous, knowable with investment,
highly solution relevant, high threat potential, unintentional and recognized.
Four (11.20, 16.10, 19.10, 20.20): future related, ambiguous, now unknowable, relevant, potentially
benefiting, unintentional and recognised.
Four (3.20, 7.10, 21.10, 11.10): future related, ambiguous, now unknowable, relevant, high threat
potential, unintentional and recognised.
One map (6.1) was exceptional as it contained only mono-polar non-knowledge: past-related, clear,
knowable, not relevant, potentially benefiting, unintentional and recognised.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
48
Figure 12 shows the distribution of 40 unknowns according to the gradients of the dimensions. The colours represent the
code used throughout this thesis (compare Figure 11). The specification to the right gives the concrete terms for each
presented colour and is probably more intelligible than the -2 to 2 coding.
Figure 12 shows that almost all (35) mapped unknowns were recognized and unintentional. Most
were ambiguous (31) and related to the future (26). About half were highly relevant (22), had a high
threat potential (20), were potentially benefiting (18) and were now unknowable (17). About a
quarter was not relevant (11), extended through the present (10) and was clear (9). Some were
knowable with investment (8), would be relevant (7), knowable (6), retrospectively recognized (5),
and related to the past (4). Two unknowns were neither benefiting nor had a threat potential and
one was neglected (non-)knowledge.
Several interviewees complained about the difficulty of answering the questions of the map, as they
are not used to the (double) negative formulations and said it would be “mind-fuck”. One
interviewee said that Global Change Management alumni would be generally sensitised for non-
knowledge but that it was still difficult to answer such detailed questions about non-knowledge.
In the following, reflections to each dimension from exploration (theoretical part) and interviews are
presented. The succeeding tables for each dimension are separate rows of Figure 11 with additional
information. It is presented how many of the 40 unknowns mapped in the interviews belonged to
each gradient of the specific dimension. Exemplary unknowns with their decision context illustrate
the gradients.
010 20 30 40
Recognition
Intentionality
Threat Potential
Solution Relevance
Knowability
Ambiguity
Temporality
Distribution of 40 unknowns among gradients of each dimension
-2 -1 012
recognised - retrospectively
recognised
past - present - future
clear - ambiguous
knowable - knowable with investement -
now unknowable - unknowable
not relevant - would be relevant -
would be relevant - highly relevant
potentially benefiting - not
determined - high threat potential
unintentional - neglected (non-)
knowledge - intentional
specification of gradient (left to right)
colour code for gradient as in tables
dimensions
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
49
3.3.1.1
Temporality
Temporality past-related
extends through
present
future-related
out of 40 maps
4
0
0
10
26
Examples
I am asked to verify
an event in the past.
I do not know what
has happened. Now
I have to decide if I
want to (or can) find
out. (circumstances)
I have to take a
decision in a
political situation.
I do not know the
intentions of
other actors.
(circumstances)
I have to take a
decision and
cannot ask my
team. I do not
know how they
will react.
(consequence)
“Non-knowledge can be generated by non-contemporary factors, for example, when a person cannot
know what happened thousands of years ago or what will happen in the future. While archaeology
and historical science help our understanding of the events of the past, the future-related non-
knowledge is entirely unknowable.” (Ibisch and Hobson, 2012a)
For a practical application of the map in decision situations, the future is often not generally
unknowable (compare 3.3.1.3 Knowability). It is for example commonly understood as knowable that
public transport will function (although not always as scheduled), that planned actions, e.g. elections
or events, will take place or that certain behaviour results in a certain reaction, e.g. releasing a ball
on a slope or anticipating social behaviour – e.g. as (Kahneman, 2011) describes it: “You saw that the
young woman’s hair is dark, you knew she is angry. Furthermore, what you saw extended into the
future. You sensed that this woman is about to say some very unkind words, probably in a loud and
strident voice.” Whereas this “knowledge” about the future often relates to rather close time
horizons, modelling, simulations and scenarios point to several more distant futures. They are
understood as being false but thinkable. It becomes clear from those examples that the future is
partly knowable or can be approximated with a degree of certainty. However, those transitions
between knowable and unknowable futures are smooth in reality, and even though they are largely
common sense, they differ between individuals (3.3.1.2 Ambiguity).
The category “extends through the present” has been added as in some decision situations, the non-
knowledge relates to something that is happening at the very moment of the decision or in a time
window which extends through the present and hence includes past and future as well. The
according knowledge can for example not be accessible (knowable in the decision situation) because
it happens in a different place or mind, e.g. Schroedinger´s cat or intentions of other actors. A
frequently observed form of non-knowledge extending through the present is non-knowledge how to
do something, but as soon as it is decided how it should be done, it is determined and hence
knowable.
3.3.1.2
Ambiguity
Ambiguity
clear
(unambiguous)
retrospectively
ambiguous
ambiguous
out of 40 maps
9
0
0
0
31
Examples
I have to
decide if to
certify my
product but I
I do not know if my
partner accepts the
compromise I
suggest. I thought
I have to decide if
to do a project. I
do not know
about the
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
50
do not know if
the
certification
will be
successful. This
is clear (either
certified or not
certified).
it would be either
accept or not
accept – hence
clear. But my
partner proposed
something
completely
different. Hence I
recognize
retrospectively it
was ambiguous.
security situation
in the country.
The security
situation
however will be
ambiguous as it
relates to the
project.
Some non-knowledge is ambiguous; it may vary between individuals or also within one individual.
This became clear during the interviews. Sometimes the author had assigned different classes of non-
knowledge than the interviewee did.
Clear (rather than unambiguous – which is harder to understand) and ambiguous mark the end of the
dimension. Retrospective ambiguity was inserted as a degree of ambiguity. This implies that the non-
knowledge seemed clear but was ambiguous: it is retrospective ambiguity – when you took the
decision you thought your non-knowledge would be clear but retrospectively it turned out that you
oversimplified your non-knowledge – it was actually not clear but ambiguous.
Ambiguity can either be in the issue itself or it can be in the user of the information. Even if one has
complete knowledge about aspects of a product, one still has to decide which aspects matter most.
And this can be highly ambiguous for different situations.
Dealing with ambiguity is described as “often we prefer to accept ambiguity than to try and find out
all details. The brain solves ambiguity related problems by using heuristics” (Gigerenzer, 2007a), but
also jumping to a conclusion can resolve ambiguity (Kahneman, 2011) or “the halo effect is also an
example of suppressed ambiguity: like the word bank, the adjective stubborn is ambiguous and will
be interpreted in a way that makes it coherent with the context”.
3.3.1.3
Knowability
Knowability knowable
knowable with
investment
now unknowable unknowable
out of 40 maps
6
8
0
17
9
Examples
I do not know
the
environmental
impact of my
project.
However, as it
is defined, e.g.
by an
environmental
impact
assessment, it
is knowable.
I have to decide
if to publish the
results of a
study but I do
not know (I
doubt they are
not) if the
results are
correct. If I
invested some
time I could
know.
I have to organize
a workshop and I
do not know if
the audience will
have a positive
attitude towards
the topic. It is
now unknowable
but it will be
knowable after
the workshop.
I have to organize
a public relations
event and I do
not know who
takes home the
message. This is
ambiguous and
not feasible to
measurable. It is
generally
unknowable.
“This is a key dimension of non-knowledge and is determined by differentiating between
theoretically knowable non-knowledge (the knowledge gaps to be filled in the course of the
advancement of science), and the unknowables. Things can be unknowable for various reasons. The
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
51
most obvious reason is relate to temporality (…), particularly in the context of future events.
Important forms of the unknowable are generated by patterns and dynamics within complex systems
behaviour that lead to indeterminacy or uncertainty. If potentially dangerous, we would describe
these as systemic risks.” (Ibisch and Hobson, 2012a)
The first sentence has to fit non-knowledge in decision situations, it would spell: This is a key
dimension of non-knowledge in decision making and is determined by differentiating between the
practically knowable non-knowledge (the knowledge gaps to be filled within the decision´s time
frame), the knowables with investment (I decide not to know (compare 3.3.1.6 Intentionality)
because it would delay the decision, require monetary investments, cause psychological stress etc.),
the now unknowable (or the undetermined, which will only be generated or observable for the
decision maker once the decision is taken, non-knowledge attached to the decision (we don´t know
before we try)), and the generally unknowable which is detached from the decision (undertermined,
unknowable, uncertain).
Hence the side of knowable is extended for decision situations. The side of knowable is extended
with knowable by investment and the side of unknowable distinguishes between the generally
unknowable and the now unknowable.
Regarding the term ‘knowable with investment’ the investment will usually refer to time or time
‘plus’; ‘Plus’ could be social stress or money. It is for instance knowable if one delays the decision and
does further research, tries to reach the experts/locals that know, invests in someone generating
that knowledge etc. It could also be called “unknowable without investment” because it is assumed if
someone chooses “knowable with investment”, the person did not invest because if he or she had
invested it would be known (and hence not unknown and never be analysed by the map).
Now unknowable often relates to “undetermined” non-knowledge – if it works will be knowable or
determined as soon as the decision is taken (turned into knowledge – only if you do it you will know).
The role of half-knowledge, doubt and approximation remains in the eye of the beholder.
3.3.1.4
Relevance
Relevance not relevant
would be
relevant
would be
relevant
highly relevant
out of 40 maps
11
5
0
2
22
Examples
My institution´s
policy is not to
publish negative
studies. Hence it
is not relevant if
the study results
are right or
wrong.
If I decide to
play the
lottery I do not
know if I will
win. If I could
know if I win it
would be
relevant.
I have to advise
someone. I do
not know their
political and
strategic
thinking. If I
could know it,
this would be
relevant.
I have to decide if
to change the
methodology for
an event. I do not
know if the new
method will work.
This is however
highly relevant.
“In the science of sustainability there is more knowledge about problems than there is about
solutions. The non-knowledge approach to finding solutions to problems is more relevant to
achieving sustainability than any concerns about a lack of understanding of the problems. The
definition and perception of problems is context-dependent and tends to be ambiguous” (Ibisch and
Hobson, 2012a).
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
52
In identifying a decision and asking about corresponding non-knowledge, solution relevant
knowledge is implied. However, sometimes interviewees named irrelevant non-knowledge. This
irrelevant knowledge might have a function – in terms of redundancy or because it is relevant to a
different decision point within the same process.
However, a lot of discussion took place if the “would be relevant” is on the side of relevant or the
side of not relevant – and if it exists at all and was not a trap of knowledge. Would have been
solution relevant might be retrospective and evaluative towards the taken decision.
Often decision conditions or the framing of the decision can be in a way that specific unknowns are
not relevant. Decisions can be framed in a way that with or without a potentially benefiting non-
knowledge it would be worth taking the risk, even though objectively it only made sense if it was
positive. Hence, it is not only about decision making, but also about how to formulate and perceive
the decision. Decision making starts long before that.
3.3.1.5
Threat Potential
Threat
Potential
potentially
benefiting
high threat
potential
out of 40 maps
18
0
2
0
20
Examples
My task is to
activate our
members. I follow
the public debate.
I do not know
who my target
group is. This is
potentially
benefiting.
I do not know if it
is positive or
negative. If it was
positive it would
be potentially
benefiting. If it
was negative it
had a threat
potential.
I do not know
how they will
react. Their
reaction has a
threat potential
to me (I rather
expect a
negative
reaction).
“This dimension specifically identifies the risk associated with certain non-knowledge. The
conventional definition of risk is an event that may impact a certain object, thus the severity and
probability of such an event occurring should be understood. Non-knowledge itself cannot be a risk,
but it can draw attention to events that may harm the integrity or functionality of systems. Threat
evaluation is not entirely objective and tends to be value-loaded (see below: ambiguity). In cases
where the status quo is the primary objective, the presence of a threat will increase the vulnerability
of a system. An alternative non-equilibrium approach that accepts forces of change would not
necessarily recognise these same threats as sources of danger to the system.” (Ibisch and Hobson,
2012a)
This is the only antifragile dimension on the map – the opposite of high threat potential is not low
threat potential but potentially benefiting (opportunity).
Risk in decision situations is value-loaded and cannot be determined. Its severity also has to relate to
a specific individual or group – e.g., is there a threat potential for the decision maker, the
surroundings, the organisations, or for a wide community?
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
53
3.3.1.6
Intentionality
Intentionality unintended
neglected (non-)
knowledge
intentional
out of 40 maps
35
0
0
1
4
Examples
I have to
decide if to
support a
certain event. I
do not know if
this event will
reach its goal.
Hence my
non-
knowledge is
unintentional
I have to decide if
to do a project in
a certain place. I
do not know the
exact security
situation. I could
however have
assume it would
be a major
constraint. I
neglected my
(non-)knowledge.
I have to decide if
to do something
as a surprise.
Then I do not
know how the
person reacts.
This is intentional
non-knowledge.
“Non-knowledge may prevail in situations where ´knowledge architects´ choose not to acquire
knowledge (fully recognized) thus creating a deficit. There is a wide spectrum from carelessness or
lack of interest in cognition to the active and normative rehabilitation of non-knowledge (Wehling
2007). As mentioned earlier in the chapter, the modern interpretation of the terms ignorance and
ignorant has negative connotations. Ignorance of specific sites or objects can translate into taboos in
certain cultures, and this has afforded protection to these features. On the other hand, where
knowledge is seen to threaten traditions and lifestyles, it is rejected by many and this may have
repercussions for sustainable development.“ (Ibisch and Hobson, 2012a)
The extremes of the intentionality dimensions remain intended and unintended. On the side of
intentional, “neglected (non-)knowledge” is added. This means that the decision maker has had a
short notion of (not-)knowing something but refused to look at it in more detail or forgot about it. It
refers to (non-)knowledge which was too blurry to be admitted or confessed in making the decision.
(Non-)knowledge is in bracket because this often displays a highly ambiguous gradient between
knowing and not-knowing. The same individual might claim “I knew it” in a certain situation,
defending the ability to anticipate or “I did not know it” defending the lack of looking at this in more
detail. It is closely linked with relevance and can only be seen in retrospect.
3.3.1.7
Recognition
Recognition
fully
recognized
retrospective
recognition
unrecognized
out of 40 maps
35
0
0
5
0
Examples
When I
decided to
change the
methodology
of the event I
recognized
that I did not
know if it
would work.
When I decided
to do a certain
project I did it for
the health
benefits it
provided. Back
then I did not
know about the
positive
environmental
impact.
Blindspot: no
example by
definition.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
54
„Individuals and societies are more often than not aware of their cognitive limitations. This
awareness is influenced by pre-existing knowledge (the more one knows, the more apparent is the
gap in knowledge), norms and values, or technical and physical restrictions. By identifying non-
knowledge, it is easier to make an assessment of the theoretically knowable (knowledge gap) and the
known unknowable. The unknown unknowns are blindspots. They are the cause for surprise and can
present a threat where decisions may be made based on completely false assumptions. In the
formative years of human development, virtually all non-knowledge presents itself as blindspots.
Through a process of nurture and experiential learning, together with the help of science, the
number of blindspots diminishes. When managing for sustainability the question arises whether
effort should be spent on blindspotting, in order to detect and eliminate blindspots, or whether it is
better to adopt a position of acceptance that they exist and should be factored into all strategies for
sustainable existence. The active management for blindspots both at individual as well as societal
levels is part of governance and of political relevance. Typically, in totalitarian regimes the common
practice is to restrict public access to certain knowledge and in so doing exert greater control.”
(Ibisch and Hobson, 2012a)
Retrospective recognition is added. Even though retrospectively recognized non-knowledge did not
influence the decision, it is however often reported from decision situations (when a blindspot was
spotted). There is only a non-knowledge map if the unknown was recognized, or retrospectively
recognized. There are always an uncountable number of maps concerning the unrecognized non-
knowledge.
3.3.2
How Did You Take the Decision?
The question aimed at finding heuristics or other decision strategies. There are 39 decisions.
Interviewees said explicitly in 16 cases that they decided intuitively (compare annex VIII.3 for a
definition of terms and VIII.4 for coding of keywords). In 13 cases they said they had analytically
thought about it, in 11 cases it was a rational decision. Eight decisions were based on discussion and
five on an investigative approach. Three decisions each were said to be based on experience, limits
(circumstances) and directives. Two decisions each were said to be based on compromise (middle-
ground), personal research, sense or plausibility, some facts, consultation and in two situations the
interviewee did not know what it was based on. For one decision each it was said the decision was
taken hopefully, because I wanted it, because it was special, by expert intuition, on evidence, on my
values, on familiarity, carelessness, an experiment (try it, feel it), evaluation and feedback, following
the debate, the precautionary principle or generally informed.
Interviewees themselves explicitly gave between one and six (average 3.6) keywords for how they
took the decision. The idiosyncrasy question found 34 decisions to be based on the mind-set of the
decision maker and 29 on their values. Based on the interviewee´s remarks the author noted
additional four to 15 (average 10.2) implicit keywords.
The implicit keywords on how the decision was taken included 34 decisions that were informed by or
based on facts, 31 on sense or plausibility, 28 to 20 on experience, consultation, limits, personal
research, analytical thinking, hope and trade-off compromises, 19 to 10 on familiarity, discussion,
evidence, following a debate, declared intention (because I wanted it) and an investigative approach.
All other keywords were implied nine to zero times.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
55
All decisions were informed by system 1 and system 2. One decision was not based on information
that is out there. Five decisions were taken without an extended peer community. Seven decisions
were taken without experience. Two were taken without subjectivity. 17 decisions were not based
on “something”. Four decisions were directed. 10 decisions were influenced by limits.
“I took the decision strategically” was implied in half (19) of the decisions. A strategic decision should
usually reach several goals in a complex situation. However, this category was analysed separately
from the others and only later on (for the statistical data analysis) integrated into “informed by
thinking”.
Some keywords that were suggested in the questionnaire (“it was clear”, “I wanted it”, rational
analysis, unhappy about alternatives, precautionary principle) were mentioned as explicit keywords.
The suggestions “to satisfy someone else, it meant progress, it was part of a package, not to be
blamed later, muddle-through, “I always do it like this”, “I was taught to decide like this”, “I decided
randomly or by chance” were not suggested by the interviewees. Some could implicitly be assumed
but were not analysed in more detail.
3.3.3
Non-knowledge Literacy
In the following paragraphs, the principles suggested in 2.1.6 are related to the interviews. Various
interviewees were working with more or less institutionalised frameworks of adaptive, integrative
and flexible management. They seemed to be implicitly (although not explicitly) familiar with no or
low regret strategies and even more so with robust or resilient, participatory, proactive, precautious
and preventive strategies.
From the proposed aspects of non-knowledge literacy in the literature, some seem to be hard-wired,
others seemed to be grounded in a mind-set which most Global Change Managers appear to have
and others are not represented. A list which evaluates the principles of non-knowledge literacy
(2.1.6) in conjunction with interview behaviour can be found in annex VIII.1. The results of this
metasystemic analysis are synthesised in the following.
• Naturally there, hard-wired: prepared, muddling-through and guidance by the
circumstances, basic distinction between important and unimportant unknowns, basic trust
in intuition, shrinking your boundaries of consideration
• Mind-set (individual and institution), principles and strategies:
Largely observable in interviewees: positive attitude towards the unknown, use of no or low
regret strategies, robust or resilient, participatory, proactive, diverse, systemic, scenario-
based, precautious and preventive strategies, adaptive management, flexible and integrative
approaches, principles of problem solving, avoidance of absolute terms, take responsibility,
allow intuition, decisive action
Observable in a few interviewees and not to the full extent: transdisciplinary approaches,
doubting of immediate causes, limiting the interaction with the unknown, avoidance of
focusing.
Not observable: humility, non-knowledge based and metasystemic management, less certain
worldview, antifragility, communicate uncertainty in politics, econics, challenge your
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
56
routines, maintain redundancy, recognising unknowns timely (to allow acting accordingly),
more sophisticated distinction between important and unimportant unknowns (can be done
with the relevance dimension of the map), avoid exposure to small probabilities in certain
domains, undirected (bias-conscious) tinkering, omission, set boundaries, allow intuition
even if it ignores information, delay hypothesising
• Should be put out for discussion, trial or scientific investigation: injecting unknowns to
stabilise the system, making many small mistakes and allowing to break what needs to be
broken.
• For econics and redundancies, just the terms would need to be spelled out, so that their use
is conscious. Then it could be used as argumentative strategies and overarching concepts.
3.3.4
Post-normal Science Diagram Inspired Questions
Figure 13 shows the result of the post-normal science diagram inspired questions. The green bars show the distribution
among low, medium and high for the respective question. The blue bars give the average of the preceding green bars and
answer the respective overarching question.
The majority (23-30) of the decisions had medium decision stakes, medium environment volatility
and were taken by individuals with medium idiosyncrasy (compare Figure 13). 11 and 12 decisions,
respectively, were taken in environments with low volatility and by individuals with a high
idiosyncrasy. About five decisions had low or high decision stakes and were taken by individuals with
low idiosyncrasy. Just one decision was taken in a highly volatile environment. The environment
volatility tended to be low and the idiosyncrasy to be high.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
57
Formative for the decision stakes were especially the high impact over time (25 decisions), people
influenced across the own institution or community (medium, 30) and a medium investment of time
and money (20). The volatility of the environment was largely influenced by a low social volatility (31)
and around half of the decisions with a topic that develops but where changes remain calculable
(medium, 22) and in institutions that are large (19) or medium sized (19). The idiosyncrasy assessed
for the decisions was in 25 cases taken with a strong mind-set, most interviewees could answer the
questions with few exceptions (medium, 18) and had some of their values influence their decision
(medium, 17).
Figure 14 shows all decisions on the post-normal science diagram. Green dots or background means low, yellow medium
and red high idiosyncrasy. The colours could be extrapolated globe-like into space. The red line (two dimensional only)
shows that decision stakes vary less among decisions than the volatility of the environment.
In the two-dimensional representation, five decisions are in the green part of the post-normal
science diagram (compare Figure 14). 17 decisions are in the yellow and 15 decisions in the red part
and two are out of the diagram.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
58
In the three dimensional representation (colour coding extrapolated), two are in the green part,
seven in the yellow part. At least two would be outside the red area.
The majority of interviewees answered the questions without difficulty, but a few issues emerged.
In terms of influence, one interviewee explained "That´s difficult because I can´t really tell what
would have happened. I decided to do this and the outcome went as I thought it would go. It could
probably – I mean this doesn´t mean if I had not decided this, it would not have come to the same
outcome. I can´t really say that this is a consequences of my decision". Others argued systemically:
“everything is interconnected – of course my decision influences a wide global community“. For the
decision investment, one interviewee questioned if it would not include other aspects of investment
such as the effort made. The first option was usually chosen by interviewees for instantaneous
decisions without any monetary investment but in one case it was an investment of several thousand
US$, as it always related to the decision context (environment). On the other side, major decisions
with large investment for a project might be very small in comparison to most issues. Difficulties
emerged for interviewees who worked in communications where events had a strong short term
impact but aimed at a long term impact too.
The definition of the topic was left to the interviewee so that the scope of this question varied, e.g.
from climate change to compliance with standards. One interviewee agreed with working in a large
institution, but said that changes happen on a daily basis, but for all others these categories seemed
to fit. One interviewee works for an NGO where the social environment consists of volunteers and
hence the relation to boss and colleagues is very volatile. One other interviewee said “I can foresee
my colleague´s reactions but they are not coherent, my component is stable, it is well-organized
intentional necessary volatility. I don´t know. It is a tough one. It is there. But it is their job. It is how
we - how it works. The way the institution works. If it is volatility it is necessary volatility, because all
decisions are made this way, in this institution.” In this quotation the understanding of volatility is
different. The interviewee could be convinced that as it is foreseeable, it is a high form of stability
and not volatility.
The initial formulation of the value and mind-set questions caused confusion amongst interviewees
(3.3.6) so that it was adapted for all interviews. One interviewee distinguished between personal and
professional values and was told to give personal values. One interviewee illustrated: “I used my
mind set and my qualifications and all that stuff. It was a subject matter it required understanding of
the technical context.”
3.3.5
Exemplary Interview
One exemplary interview is presented (anonymised and alienated) with all given and assigned
variables. This interview was chosen because the unknown is very generic. There is only one decision.
Evidence
Yes, my decisions are 100%, always based on evidence - but probably not as you define evidence.
Please explain a global change related decision you took at work.
I coach a decision maker. I did an analysis of his team some time ago. Now I had to decide how to
present the results of this analysis to the whole team. They are very cognitive and I thought I had to
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
59
do a classical presentation (but I did not feel quite well with this). Yesterday I decided to use a
different method. Preliminary, not yet implemented, complementary.
How did you take the decision?
I talked about my uncomfortable feeling with the classical method with a group (discussion,
consultation, supported). Then I had the idea to try the new method so I tried the new method with
my group (investigative approach, which gave me evidence how and that it works). I thought the
new method through, and wanted to understand it (some facts, personal research). Then I double-
checked my new method with two experts.
Assigned: bold (see above) and sense, based on feedback from similar situations, experience, expert
intuition, hopeful, because it was special, my values, because I was familiar with the method, because
I wanted it, limits (unhappy with alternatives).
Decision conditions
This is an important part of my job. A requirement from the coaching. Because I had the idea. I
could influence the alternatives and took all the time I needed (even though it was not much time).
What did you know when you took the decision?
Assigned: circumstances (6 - six knowns were given by the interviewee), consequences (2),
complementary (1)
What did you not know when you took the decision?
There was nothing relevant I did not know.
Assigned: circumstances (4), consequences (1), complementary (0)
Mapped non-knowledge
Now as we are talking about it I realize that I did not know what will change, “change is the only
constant”.
-2
-1
0
1
2
Temporality past-related
extends through
present
future-related
Ambiguity
clear
(unambiguous)
retrospectively
ambiguous
ambiguous
Knowability knowable
knowable with
investment
now unknowable unknowable
Decision
Relevance*
not decision
relevant
would be decision
relevant
highly decision
relevant
Threat
Potential
potentially
benefiting
high threat
potential
Intentionality unintended
neglected
(non)knowledge
intentional
Recognition fully recognized
retrospective
recognition
unrecognized
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
60
*Was or is it relevant what will change? “It is completely irrelevant. It is relevant because it will
change our interaction but it is irrelevant for me because it - since it is everything - I don´t care. I only
know it will change something. It is always like that. It is a law of nature, I think, and therefore it is
not relevant. I just know that I have to adapt. So it is not relevant. But I know we have the
preconditions to react to the change.“
Of what type was the non-knowledge?
Assigned: future (temporality)
Decision Stakes
- My decision influenced people across a wide (global) community (3)
- The decision required no or little time and money (1)
- The impacts of my decision will be observable for more than a year (3)
Decision environment uncertainty (VOLATILITY)
- The topic (process guidance) is prone to abrupt large changes (3)
- My institution is large, structured and has been in place for a long time (e.g. ministry, large
company) (1), (but I do not agree with “changes are rare”)
- The relation to my boss and colleagues is very stable, I can rely on them and anticipate their
reactions (1)
Idiosyncrasy
- My values probably had some influence on my decision (2)
- My mind-set was the main reason for my decision (3)
- The interviewee got into thinking about non-knowledge easily (3)
While deciding I was content. Today I am pretty sure I took a good decision, but could still change
this. I think this is based on my ability to make good decisions, but my wording would be more
humble.
3.3.6
Exemplary Decisions
Six particularly interesting or illustrative decisions or their combination with unknowns were marked
in the spreadsheet and are presented in this section.
A. I work in public relations and making such decisions is at the core of my job. But this time I had
the idea to do a surprise. Now I had to decide whether to do it. I did not know how my bosses
would react as it would be a surprise for them. However, I felt it was potentially benefiting.
B. I work in an association that is in favour of a topic, so it is our purpose to publish positive
information about this. In this case, there was something that was against our topic so it is our
directive not to publish it. I was, however, not sure if the information was right or wrong (value
non-knowledge), but this was not relevant to the solution as we took the decision according to
the directive and not according to the real information.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
61
C. Before I studied GCM I worked in a project and I had to decide about a small budget. I decided
to do something which has health benefits. Back then, I did not know it also had
environmental benefits (retrospectively recognised). (here the values of the interviewee
changed so that now something could be perceived as knowledge that was hidden from
consciousness before (a blindspot))
D. I had to decide which partners to choose for a project (what exactly). We already had some
contact to potential partners and there was not much time (limits). I did not try very hard to
get in contact with other potential partners so finally I decided with stick to the ones I was
familiar with. However, I did not know much more about them than what we had exchanged
in a few emails - that was basically nothing.
E. We had to decide where to put 20 dams. I had all the knowledge about good locations from a
spatial analysis. However, I had to decide together with locals. I did not know their internal
power structures. So even though my information and my knowledge from the analysis were
very firm, I could not quite understand the decision process.
F. In my studies I learnt a lot about something which is generally underrepresented in the public
debate. In the decision, I felt this something would relate a lot to the document I was writing
and we were publishing. So I had to decide if to take this aspect in. I did some personal
research on the issue. I could however not find concrete evidence but I still felt it was crucial,
as it was said to be so important in my studies (mind-set) and it really became my values. I
decided to write about it and could convince my team. „Mostly - I hope - I guess“ that my
decisions at work are based on evidence. I think this unknown is knowable.
3.3.7
Discursive Interviews
The five most discursive interviews are presented in this section. Contents and comments relevant
for discussion were clustered according to by how many they interviewees were treated or given.
The five interviewees are representative in terms of nationalities, employers and in knowing the
researcher, not in terms of gender. The interviewees tended to have more total work experience, be
more interested in the results and often scored a three in interview difficulty. The interviews took on
average 60 minutes. More meta-level and ambiguous but less relevant non-knowledge was reported.
The reported decisions were taken more intuitively than rationally, the interviewees had finished
their studies faster and had held their position for longer (compared to the average, the remainder of
interviewees and a random number of five other interviewees – these calculations are not published
to maintain the anonymity of the five interviewees). In most of what those five interviewees said
they were exceptional, if not, it is indicated in brackets.
These five interviewees were mostly able to look beyond the mere question and often answered to
more than asked and gave additional examples. They frequently changed perspective, went to the
meta-level and reflected upon issues, e.g. with counterfactual reasoning.
Four interviewees complained about or criticised some questions and options, suggested
improvements and additions, refused to answer etc. Finally, all interviewees could be convinced to
answer and find a second decision and the disputes were settled. They were promised the analysis
would be quantitative and qualitative.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
62
• “Sorry. I have to destroy your categories”,
• “Honestly. I think those questions are not very well suited to the kind of decision I
described”,
• “Can I not like not answer that question? Because it is not applicable. (I would be very
happy if you answered). I don´t want to interfere with your scientific way. I don´t like
surveys much for that reason.”
• “I can´t think of a decision that makes sense in the questions you were asking.”
• “can´t there be a third option like - okay…that´s fine.”
• “You are asking the wrong questions, I don’t understand”
• “Was this decision relevant? … You need some thresholds. It´s not so much compared to
why the climate is changing.”
• Assuming the purpose of the interview: “I don’t have an example for you. I think I
understand where you want to go. Delete my previous answers. My example does not
fit. I took an evidence based decision, you are looking for an intuitive decision.”
• “I think your topic is very hard. Very difficult to ask good questions”,
• “Can you read the question again. I don’t understand. Can you read the question again.”
• Two heavily supported the change of the mind-set (and value) question, so that the
other three had the new version where this question was adapted.
o “I think you need, it´s maybe too late, your first and last criteria are opposing each
other. The values and trying to stay objective. Even if my values influenced my
decision a lot I am still objective. You were not like carried away by your values? … it
gives the feeling that - I personally would say I have strong values and I usually try to
not lose my values in personal or work decisions. I am a strong believer in global
human values. You shouldn’t lose your humanity. The term objective is not needed in
that concept. I don’t want to give answer number three. Because I am tending to go
away from my values in order to take an objective decision. I am highly driven by my
values but of course I am objective and fair or try to be objective and fair as I am. I
don’t see these as opposing concepts, having values and or being objective. I was
struggling between 1 and 3. I would still say number 1. I am strongly driven by values.
I am still objective.”
o “If I think your question is asked in the wrong way and I would like to say one and
three at the same time, how can you tell why I say rather one or rather three, if you
are forcing me to decide for one.”
o “I want to say 1 and 3. There is no multiple choice in here, right? But 1 – it´s
falsifying. I don’t think this is good.”
o “The options are strange”,
o “Maybe mental models is a better term”
Four of those five interviewees conveyed the impression they had explicit mind-sets or strong values
which they used to evaluate the questions and explain their answers. The mind-sets seemed to be
either provided by the institution they were working in and they had internalised so much or by the
study programme itself.
Three seemed to be experts in their field.
Three of them were rather critical, questioning and defiant.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
63
Three interviewees expressed their understanding of non-knowledge or shared their reflections
about the boundaries between knowledge and non-knowledge:
• “More knowledge was possible but not all of it. You cannot predict what will finally happen.”
• “we call that bounded unknown. I did not know but I could kind of make an estimate about
the worst case”,
• “The knowledge I acquired has very little history. If I compare it to knowledge of nature and
the totality of non-knowledge. I feel my knowledge is nothing compared to the non-
knowledge of nature. ”
• “Of course there is not proof. It is suspicion. Like a half-knowledge, I have no idea what my
basis is, I assume what their basis is, the way people argue and some of the things they
admit.”
• “All of that was pretty much knowledge you said? No, non-knowledge, that´s my
assumptions. Part of my non-knowledge is will the future really go in that direction. The
other non-knowledge is in that direction but I would say I am 90% sure but I have no
evidence for this except discourse. The way, I don´t know how to talk about that. It´s just like
the way they talk. The things they complain about. That in a sense is pretty much evidence. It
is not complete knowledge. This is one actor, where I have the most knowledge.”
• “Knowledge in totality, there are two parts one is non-knowledge and one is human acquired
knowledge. Leaving all other knowledge as non-knowledge. We can say the rest of human
acquired knowledge is non-knowledge. The knowledge is within the system it is in the
lithosphere, in every other species, but the limited knowledge of this non-knowledge - in
decision making we only have the human acquired knowledge part play. We ignore that non-
knowledge part. This is the problem.”
• “We have a nature of taking risk, of gambling. Chernobyl and Fukushima show this. The
problem is that logic advises us against these practices but experience tells us that risk comes
with high return benefits - so our problem is that few of us only take care to phase out… it is
the nature of humans to take risks and gamble.”
Two proposed their own heuristics
• “precautionary principle”,
• “I always analyse the why, what and how”,
• “double-check”, “I went back and forth several times”,
Two interviewees expressed how positive they thought this topic was and expressed their thanks in
advance and during the interview “thank you, really good question. Really good interview, because
you are helping me to get clarity on something that actually happened in the decision.”, “perfect.
That´s a great question!”.
Two criticised one size fits all projects.
Two asked “Intended and recognised by whom?” which indicates incomprehension of the
dimensions exact reference.
Two said the same aspect for knowledge and non-knowledge.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
64
For two it was extremely difficult to find a decision and two were expressing doubt if their decisions
would fit the questions: “Now that I know these questions I could think about another decision” or
“...we took a stupid example!” (that happened in other interviews as well that people expressed
regret or embarrassment about the decision examples they gave). One interviewee ordered to
change the suggested questions several times: “Write something else down”, “Delete my previous
answers.” (It occasionally happened that interviewees changed their proposed decision after a few
sentences).
Two said “I don´t care” about what they did not know.
Two related everything to a bigger picture and lost focus between the question and the decision
point/process.
• “Related to the impact this has on the issue, the investment of time and money was
mosquito shit”,
• “but this decision is nothing compared to the real problems we have in the world”
• “While my example is so small there is almost no investment, quite a lot for me, but nothing
compared to major global question”
One interviewee (expert) regularly lost the decision point. The according decision process was still
ongoing and there was a continuous skipping back to the larger decision that was taken on a
different hierarchical level. "I don´t feel I decide". “The questions you asked are very much focused
on a single person´s decision making. I wonder if it´s possible to talk of group decision making
because that is in the end - then it gets a completely different dimension. And basically all I do is
group decision making. ... It´s hard to say because sometimes I don´t feel like I am deciding
something. I feel like it´s hard to explain but because it´s this group decision making processes
oftentimes the “I” is hard to find. Like I don´t feel like I, I lose my individual self because then the
individual is maybe my institution, which is like a whole bunch. It is often as small as I get. And
sometimes a group even larger is taking part in this group decision making. We are the smallest unit
taking part in some other group decision processes. Quite rare that I myself as an individual take
place in a decision making process. I am like a cell within the organism. It´s sometimes hard to
differentiate between individual and group – this would be something for your doctoral thesis.”
One interviewee reflected upon the values and mind-set which were very close to the values and
mind-set of the institution, and it was asked if with different values and mind-set he or she would
actually work in another institution (what came first, the chicken or the egg).
One interviewee had difficulties in understanding various questions “I don’t understand. I don’t
understand.”, “I´m confused now. Can you repeat?”, “Can you read the question again.”
One interviewee expressed how useless it felt to analyse non-knowledge in such a clear decision and
mentioned something completely off-topic “I didn’t know what the weather was in Taipei”.
One interviewee emphasised: “Decision making today is not the same. I am not the same guy as
before I studied. I have become super different from who I was.” and offered decisions from work
before GCM: “Unknowingly I did good”, “now it can be rationalised, later on, but it was intuitive“.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
65
3.3.8
Remainder of Information from Interviews
Results and hypotheses from the theoretical part and those that emerged during the interviews are
presented in the following section. Not in this section are results to the questions “how did you
decide”, “what did you not know”, the post-normal science diagram questions and the non-
knowledge map which were presented before.
General remarks by and behaviour of the interviewees
Some interviewees were in need for discussion after the interview. They got upset or saw the
interview with its detailed questions about something so ordinary as almost pointless. The perceived
triviality of the own decision was articulated several times by the interviewees, e.g. "so many
decisions do not matter, I just take them, it does not have to be the best decision".
For various questions, interviewees asked for definitions of terms (knowable, unknowable, evidence
etc.). They were told to define the terms themselves and use their understanding of it. In some cases
they shared their understanding.
A few interviewees answered to more than asked and had difficulties to focus when they were asked
for the relevance of the unknown to the decision point. A lack of focus was also observed when
interviewees (in two cases) regularly lost their decision point and related their non-knowledge to the
decision process. They were then reminded of the specific decision point.
In some cases, interviewees answered to less than asked or were anchored by early options, such as
“When you took the decision, did you feel worried (1), happy (2), or were you…” – answer “happy”
was so strong the third option was not considered, even if later read to the interviewee.
Interviewees report decision situations and associated non-knowledge after some time has passed.
Hence, they will most probably be biased by hindsight. This retrospective distortion cannot be
avoided (in a self-experiment “real-time tracking of a decision” even real time distortion was
observed); however, it can be used, e.g. if decision makers become aware of the fact that often they
had “intentional ambiguity/retrospective intentionality” – this means they often ignored non-
knowledge they partially knew; they could become more aware of this in future decision situations.
3.3.8.1
Evidence Question
Eight interviewees said their general decisions at work were based on evidence, whereas three said
they were not and ten said they were sometimes. Many laughed about this question, some asked for
a definition of evidence and some shared their thoughts of what evidence is. Answers for ‘yes’
included “Yes. Everything is based on evidence”, for ‘sometimes’ included “if possible always”,
“depends on which decision”, “sometimes, but it is not the only driver” and for ‘no’ “my general
decisions are not. It is more experience”, “no, not all of them”, “mostly not. The reason why we make
political and strategic routine decisions is based on evidence but not the decision itself”.
3.3.8.2
Decision-related Questions
Some potential variables appeared during the interviews and were included in the spreadsheet. Not
all of them are based on the used literature. Some are based on common sense.
Please explain a global change related decision you took at work
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
66
The majority of the interviewees did not perceive their work decisions as “global change related”.
They argued “I am not a global change player“, or in written communication before the interview
“I´m afraid you are overestimating my position ;-)”. Hence the question was usually reformulated as
“Please explain a decision you took at work”. However, even this retrieval of a random decision
situation from memory proved difficult for seven of 21 interviewees, three of whom did not know
before the interview that they had to give a concrete example of the decision. All others were asked
in the scheduling email to think about two decisions they took at work. For half of the interviewees it
was easy to describe two decision situations. Retrieval of the second decision was usually easier, but
occasionally still difficult. In a few cases where the interviewee´s understanding of the issue was
good and decision retrieval easy, the author suggested a specific decision constellation (it could e.g.
be a decision like…, did you ever take a decision where you did not know something that is
knowable) to the interviewee.
Some interviewees expressed concern about whether their decision was useful to the author. One
interviewee refused to explain a second decision because the decisions would all be similar and in
fact, the second mapped unknown aspect produced exactly the same map. Figure 15 shows that the
first and second decisions are often close but not very close and in some cases even opposite.
28 decisions were taken alone; 10 of those with support from someone else. 11 decisions were taken
by a group. 21 decisions were final and 11 preliminary (used like tentative), for the remaining seven
it was not known. 27 of those decisions have been implemented and six had started implementation
whereas for three the implementation had not yet started. The remaining decisions implementation
status remains unknown. Preliminary decisions can later be changed by someone else or the decision
maker, as for example adaptive management allows.
The decision task could be described by the three mutually exclusive categories “how to”, “if to” and
“what exactly”, e.g. “I have to decide how to deal with a certain stakeholder group”, “I have to
decide if to support this project” and “I have to decide for a concrete set of strategies, what or which
strategies exactly to choose” (it could also be a when exactly). Half of the decision points were
described as “if to”, nine and twelve respectively were “how to” and “what exactly” decisions.
There were 27 complementary decisions, five novel decisions and six decisions each which were
exceptional and political. Six decision tasks were purely organisational whereas 11 were content
related. 19 decisions were taken strategically.
Decision time
Half (18) of the decisions were taken in limited time. A third (13) of the decisions was taken without a
time limit and in five decisions, the interviewee was decided right away (which means, as soon as the
decision point materialised). In three decisions the decision maker was decided right away but took
some time to check the decision.
Decision Conditions
Half (18) of the decisions had to be taken because it was part of the job. Twelve decisions were the
core of the decision maker´s job (e.g. “I am working for an NGO in a political setting it is the core of
my job to make strategic decisions”, or, if one is in charge of public relations it is the core of the ones
job to respond to evolving situations) and eight decisions constituted an important part of the
decision maker´s job. One decision was taken that was not part of the job.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
67
24 decisions were a requirement, 15 a response to an evolving situation, 14 an idea, 11 an
opportunity and seven to three each were requested, due to a perceived insufficiency, a
precondition for something else or a directive.
In 29 decisions the interviewee could influence the alternatives and in an additional two situations a
little. In the remaining eight decisions the interviewee had no influence on the alternatives.
What did you know when you took the decision?
Interviewees explicitly said they knew e.g. there is not much at risk, we would lose, the project would
not fail, we have to spend the money, the situation is not stable, what I learned in my studies, they
had support or we have never done this. These knowns they expressed e.g. as I knew, I was very
clear, I was aware, I trust, I was pretty confident, I had an idea of or it was some impulse knowledge.
One interviewee commented that “this is a very general question”. For a list of knowns see annex
VIII.8.
In 37 decision situations interviewees mentioned at least one known aspect about the
circumstances. In 13 and 14, respectively, they said what they knew about the consequences and
shared their complementary knowns. Only two commented on the importance of what they knew (it
is necessary, it is super important). Circumstance knowns were illustrated the most. Interviewees
gave a total of 90 aspects of their knowledge. Two interviewees gave six aspects of their
circumstance knowledge but most gave one, two or three. In contrast to this, 30 and 20 aspects of
consequence and complementary knowledge were given. They gave up to four but usually one or
two knowns for each decision situation. For the average decision, 3.6 known aspects were given.
3.3.8.3
Evaluative Closing Questions
Most interviewees were sure they took a good decision (27) and think it was based on their ability to
make good decisions (24). About half to a third say they were happy with their decision (17) or aware
they could not know and just had to decide (13). In nine decisions the decision makers were worried
about their decision, and in six cases each, they were not or not very sure if their decision was good
and thought it were based on chance. In three and two cases, respectively, they did not want to
decide for either ability or chance but would rather choose ability over chance (3) or chance over
ability (2). One of those illustrated: “It might be based on my knowledge but not my ability to make
good decisions, this decision was an easy one for me, it was black and white, not so ambiguous, other
decisions are more grey, there is more morality, more at stake“. Most interviewees laughed about
the question if it was based on their ability to make good decisions or if it was by chance. They
requested more humble formulations. Many said “the first” and appeared to avoid saying “ability”.
3.3.9
Statistical Analysis
Results of the statistical analysis only provided few hypotheses with low validity that are displayed
here. The complete results and hypotheses can be found in annex VIII.5.
Figure 15 and Figure 16 show a wide scatter of dots. Dots represent variables of decisions in
conjunction with their mapped unknowns. In decisions 5 and 12, the decision is the same and only
the mapped unknown changed. Decision pairs from the same interviewee do not stick in the same
place but are often considerably close. For 5.11 and 5.12 as well as 12.11 and 12.22, the dots usually
stick close. In a few cases, decisions are almost opposite.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
68
Figure 15 shows the results of a PCA with the data set Cleaned Values II that has a variance of only 12%. Dots are given
numbers, the first digit is the interview, the second the decision and the third the unknown: 5.12 is the data set for the
second unknown of the first decision in the fifth interview. Groups for how to (dark blue), if to (black) and what exactly
(light blue) decisions are displayed. Biplots are too close to be sensibly evaluated. Decision pairs are usually close.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
69
Figure 16 shows the result of a PCA with the seven dimensions of non-knowledge for 40 maps. The first and second
components explain 26% and 22% of the variables respectively. This PCA suggests that (1) Knowability (NM_Kno) and
Ambiguity (NM_Amb) are independent from relevance (NM_Rel), intentionality (NM_Int) and recognition (NM_Rec), (2)
unknowable (NM_Kno) and ambiguous (NM_Amb) non-knowledge tends to have a high threat potential (NM_Thr) and
tends to be related to the future (NM_Tem) and (3) unintentional (NM_Int) non-knowledge seems to be relevant
(NM_Rel) and hence intentional non-knowledge seems to be irrelevant. Dots represent maps but no additional pattern
could be observed so that no numbers are displayed.
The within-group analysis of the data used for Figure 16 (compare annex VIII.5) suggests that there
are no large differences in the mapped non-knowledge of differently volatile environments. The
between group analysis, where the first component explains 75% of variance and the second the
remaining 25%, suggests that links in terms of environment volatility seem to be close links between
• Relevance (NM_Rel) and Knowability (NM_Kno),
• Intentionality (NM_Int) and Recognition (NM_Rec) and
• Temporality (NM_Temp) and Ambiguity (NM_Amb) that are inversely linked to Threat
Potential (NM_Thr).
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
70
Analysis of the data used for Figure 16 grouped for circumstance (23), consequence (13) and
complementary (3) non-knowledge (annex VIII.5) suggests that the more decisions in a group the
closer it is to the centre of the biplots. This essentially means that there appears to be no relevance
in this. Grouped according to total work experience, the decisions taken by more experienced
individuals are closer to the centre of the biplots, which suggests that they are less strongly
explainable by strong dimensions.
3.4 Results from Empirical Part
Few apparent relations were found between the variables
Elements of non-knowledge literacy used by the interviewees
The non-knowledge map can be applied to closely defined decision unknowns
The non-knowledge map needs further modification to be comprehensible and useful
Knowns and unknowns were about circumstances, consequences or were
complementary
4 Release Phase: Discussion
The theoretical and the empirical part are discussed in this section. The mapped unknowns, the non-
knowledge map and the term non-knowledge literacy are presented in separate chapters at the end.
4.1 Discussion of Theoretical Part
Literature, the work with the non-knowledge map and the adapted post-normal science diagram are
discussed in this section.
4.1.1
Literature
The small choice of literature demonstrated clear advantages. Across disciplines, an understanding
of terms and concepts could be generated. Various concepts could be applied that facilitated
understanding of the situations and could be linked. For all analysed concepts, the literature was
broad enough to include a large number of sources and perspectives. There was enough food for
thought.
Limitations of a small sample are that there are possibly more complete sources and maybe better
concepts. Some variables might have proven suitable or unsuitable for decision analysis already. A
larger number of sources might have led to information overload as well. Such information overload
usually leads to topics being narrowed down extremely. Heavy reductionism on the side of the
literature allowed a broad and metasystemic investigation of a large, complex and real-world related
topic.
Especially for the term non-knowledge literacy, the choice of literature was too small to explore a
significant fraction of the term. Of course, one exploration is not enough. Some principles from
scientific (Kundzewicz et al., 2007; Barker et al., 2007) and historic (Giles, 1910; von Clausewitz, 1832)
dealing with the unknown gave a preview that there must be a lot more literature available on
competent handling of non-knowledge. Management and work related literature might draw
together such knowledge. Investigating literature across scientific disciplines and practice would
provide more insights on non-knowledge literacy.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
71
The popular scientific works consulted were written by renowned scientists. They might provide
even more secure knowledge than purely scientific papers and metastudies. This is suggested e.g. by
Bauer´s knowledge filter, as cited in Witte et al. (2008).
It can be summarised that the literature was selected by simple heuristics: familiarity and availability
(Ibisch and Hobson), recommendations (from someone familiar) and reputation (Taleb, Kahneman,
Gigerenzer). The overarching heuristic was inherent interest.
Kahneman (2011) suggests eliminating redundancy from information sources. This was not done
and might cause biases, as e.g. Kahnemann and Taleb cite each other and several sources from
Ibisch, Hobson and Gigerenzer were consulted. This might give the impression that there is more
evidence and agreement than there is.
A limitation of the literature review was the strong mind-sets of the authors, which became
especially explicit with Taleb (2008, 2012). As far as possible, idiosyncratic statements were
distinguished from scientific results. However, research into post-normal science – also this thesis –
hints to the omnipresence of idiosyncrasies and rather suggests working with them.
4.1.2
Exploring the Non-knowledge Map (Compass and Table Format)
Understanding the non-knowledge map was a lot more difficult than initially thought and tinkering
with and discussing it during several weeks was absolutely necessary. The map comprises very
different dimensions and was not applied before this thesis. It was not elaborated for decisions.
The map was elaborated for sustainability. However, it remained unclear why this limit of
“sustainability related” was set.
As the author has no work experience, the chosen decisions for tinkering were from real life and this
was surprisingly far from the decisions analysed later, especially in how the decision was taken.
4.1.3
Discussion of Adapted Post-normal Science Diagram
If applied to a decision, the post-normal science diagram investigates the circumstances and the
consequences of the decision. Systems uncertainty represents in which system (circumstances) the
decision intervenes. The decision stakes represent the influence (consequences) on the system. The
proposed third dimension asks the idiosyncrasies or peculiarities of the acting system. This would be
complementary. It might not be a coincidence that the categories found for (un)knowns also apply to
this post-normal science diagram. If they are functional, as suggested here, the three-dimensional
post-normal science diagram might be an even more comprehensive tool. If similar categories
already exist remains unknown in this thesis.
Some decisions were located outside the red realm of post-normal science in the two dimensional
version. In the three-dimensional version, highly idiosyncratic individuals – such as many Global
Change Managers, would often be outside the red cover of post-normal science. This would have to
be solved.
4.2 Discussion of Empirical Part
This section discusses methods and results of the empirical part. Methods include design and use of
the questionnaire, the interviewees and the statistical analysis. In the results section, exemplary
decisions and interviews, as well as answers to questions from the questionnaire are discussed.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
72
4.2.1
Methods
This section discusses use and design of the questionnaire, the interviews, their scheduling and
clustering and coding followed by the quantitative analysis. Limitations are discussed and an outlook
for the further use of the methodology given.
The author was not trained and did not acquire knowledge on best practices of how interviews can
be described and evaluated in scientific text. This might negatively surface in the use of language and
structure. Important findings from the empirical part are presented in bullets.
4.2.1.1
Use of the Questionnaire
The interviews took on average 15 minutes longer than announced, this should have been
anticipated.
Transferring the data into the spreadsheet right after the interview was possible because the
interviews were carried out by Skype. This timely processing of information allowed to reduce all
generated information but also to extend them by opening up new columns for aspects that were
not asked explicitly but became apparent during the interview (such as if the decision was final or
preliminary, taken alone or in a group etc.).
Recording as back-up allowed listening again to insightful interviews or unclear passages. It also
created redundant, not assessed information, e.g. how long answers took on average for each
question.
It was very functional to map two decisions in one interview. The second decision was usually a lot
faster and interviewees chose a decision point more deliberately. It also provided a time buffer, so
that interviews could be stopped after the first decision.
4.2.1.2
Design of the Questionnaire
Discussion of individual questions and their development is largely covered by the chapter
development of the questionnaire in annex VIII.10.
The analysis carried out in this thesis could have gained significantly from a simple heuristic:
numbering the questions in the questionnaire.
Many interviewees called for more certainty in the questions, e.g. by asking for definitions. The
expression of pointlessness to think about the unknown and repeated lack of understanding can be
interpreted as a call for more certainty and footing– a difficulty in dealing with the uncertain.
Gigerenzer describes in his book about understanding statistics (Librevis, 2006) that certainty is an
elementary desideratum of western men and makes them weak towards delusive certitude. The
open use of the questionnaire –finally the inbuilt uncertainty– was very good and functional for the
researcher and the results, but sometimes confused the interviewee (or aroused other emotions
such as anger or impatience). Those emotions could be interpreted as low non-knowledge literacy or
simply difficulties in dealing with uncertainties. Most intellectuals today are conditioned to work with
certainties, live in stable environments and expect human-built systems (as well as natural systems)
to provide certainty and stability. Very open concepts were used in the questionnaire instead of
closed definitions. Even when the interviewees asked they were not given a definition of evidence or
knowledge and for values and mind-set they were given a whole row of descriptive terms. This left
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
73
room for different understandings and made answers between interviewees less comparable, but
the generated understandings fed into the results and discussion and amplified it. This is an
uncertainty that has to be accepted when exploring a new topic and it is called for by several authors
that society and its members shall become more accepting of the unknown and uncertain (2.1.6).
Such generated stability includes cultural advancements, such as decoupling from temperature by
clothing, weather forecasts, shelter, air-conditioning or heating but also social attainments such as
job security, insurances, multiple choice tests or human rights. But even Taleb (2012) – who
promotes volatility and argues many small collapses would prevent the big collapse, says that some
confusion would stabilise the system. And in this case, stability is positively connoted. As in ChaOrdic
systems, there is probably an overlapping area between stability and volatility that provides an
optimal space for most activities and decisions. A lot of psychological research has been carried out
on the positive benefits of stable circumstances for human wellbeing. Taleb (2008) challenges this
with a long-term perspective that is critical about stability, e.g. falling down one meter ten times
does not hurt but falling down ten meters at once might kill you.
The questionnaire was designed around the non-knowledge map, which was supposed to be the
centre of the interview but only had a minor part compared to total interview time. The surrounding
questions on decision context could have been shorter, as they also generated a lot of noise. The
intentional simplicity stated in 1 proved ambiguous, as it did not apply to most of the decision-
related questions or could not be implemented. Even though interviewees were encouraged to
formulate their decision as “I decided to X”, they gave context and explications. Furthermore, in the
post-normal science inspired questions it might have reduced information too much into prescriptive
categories that did not allow an in depth analysis and understanding.
The general information assessed give a good picture of who the interviewees were but could not be
linked to the other results as the sample is too small. Whereas a posterior classification of the
employer and a possible analysis between groups was possible, the open question about the position
held by the interviewee were dysfunctional as they impeded an ex-post sensible classification. This
was a limitation of the non-knowledge based approach.
The evidence question was slightly detached from the rest and too closed to generate useful results.
This methodological bias might be caused by the yes or no format. If there had been room to go
more in depth with this question and analyse examples it might have produced different results. At
the end of the interview, interviewees might have answered this question differently. At the
beginning it might not have had much effect and they might have forgotten it was asked. The
question was provocative and a good start as most interviewees stumbled upon it and some laughed.
If interviewees really have the association that evidence equals something good, it could be
discussed that the question was a system 1 tester: ‘can interviewees overcome their association
evidence=good?’.
Decision retrieval from memory was very difficult for a considerable number of interviewees.
However, backward asking (e.g. do you have an example for a decision with past-related non-
knowledge), which was carried out during the trial run and the tinkering with the non-knowledge
map did not prove easier. A possible solution would be a different target group, such as decision
makers or people who take more decisions because they work alone, in a small team or have
responsibility for staff. In a way it was also beneficial because many very simple decisions (e.g. which
item to buy) or very complex decisions where it was difficult to grasp concrete decision points reaped
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
74
very interesting results and had been ignored with more decision-aware interviewees. Finally, it can
be asked if the interviews should generate results that are easy to evaluate or if they should give a
more complete picture of what decisions are like.
The low awareness of concrete decision points, the ill-defined goals of reality (Klein, 1999), was
problematic in some cases, as thinking is not linear, and produced interesting additional results in
others; for instance when it led to a redefinition of the decision point or a better understanding of
the decision process. It remains important to keep in mind that this definition and reduction of
complexity through focussing on concrete decision points is necessary for scientific investigation but
not in general.
The open wording “how did you take the decision” was clear for many interviewees and aroused
questions in others. This sometimes led the author to narrowly asking “was it more rational or more
intuitive”. The open answers to this question were difficult to evaluate and cluster without
disciplinary psychological knowledge. This question was initially conceptualised as “the heuristics
question” but it did not prove useful to expose underlying heuristics and rules of thumb to the light
of the day. Exposing heuristics might need more sophisticated methods because decision makers
might also hide heuristics in the work context as they would usually be disregarded or evaluated
negatively (e.g. I take most decisions under incomplete knowledge about facts because time is short,
or when I am tired and hungry I chose the default option (Kahneman, 2011) or I take most decisions
as my boss wants it (as Gigerenzer (2007b, p. 195) describes results from a study where magistrates
decided in 92% of the cases according to how others decided). It can be imagined easily that also the
heuristics proposed through the real-life tinkering with the map “to satisfy someone else (guided by
circumstances), unhappy about alternatives (guided by circumstances), it meant progress (guided by
higher goal), it was part of a package (guided by higher goal), not to be blamed later(guided by
circumstances), I always do it like this (guided by routine), I was taught to decide like this
(prescriptively guided by rules)” influence decisions in the work place even though this could not be
demonstrated with the chosen methods (sample size, way of questions). If those and other simple
rules and heuristics are commonly used, non-knowledge literacy might (not) be needed.
Decision time categories were straightforward, only two interviewees said that they had been
decided right away but later took time until they took the decisions and had hence to decide
between the options. This question remained sparsely evaluated as it aimed to expose simple
heuristics, but no simple heuristics were found. The same applied for “why did you take the
decision” and the question for alternatives.
The knowledge question was functional in its limits. A full knowledge inventory would require
different theoretical and practical background-knowledge and might even confuse interviewee and
author as treating knowledge too long might anchor them on knowledge – which is not the point in
non-knowledge research.
The answers to the non-knowledge question suggest that it was functional to find out which non-
knowledge came to mind easily (availability) and perhaps what matters most (as it was asked, “What
do you think was the most crucial unknown?”). A certain satisficing could be observed in the author,
as she got used to four or five pieces of non-knowledge and did not push the interviewees to name
more non-knowledge. The author was maybe too understandingly of the issues to allow critically
searching for more non-knowledge. In one interview (3.3.2) the interviewee was very willing to talk
and understand and did not satisfy with suggesting all the possible unknowns but summarised after a
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
75
long dialogue into some meta-unknown: I don’t know what will change. When this topic is to be
investigated further, a more critical, searching and demanding attitude of the interviewer would be
suggested.
Retrospective distortion happened and could be observed but it did not cause any recognised
limitations. In the case of a retrospective recognition of value “back then I did not know how
important forests are”, it brought a lot of insight.
The decision which non-knowledge should be mapped usually went smoothly and was occasionally
guided to enlarge the envelope of maps from the interview series. This question fell short of a
complete picture, as it was not based on a non-knowledge based approach; the non-knowledge map
questions followed too quickly after it was decided for an aspect of non-knowledge to be mapped.
The questions from the map were very prescriptive and did not allow a really open discussion. A
more open formulation or an initial brainstorming from the interviewee of how he or she would
characterise the non-knowledge had been helpful and should be applied (in addition to the map) in
further research. Although some doubt remains if the majority of the interviewees had understood
the question “what do you know about your chosen unknown?”. A severe limit was that only one of
the unknown aspects could be mapped and hence the relation to the other questions was very thin
and the PCA could not show strong relations between variables.
The named unknowns from the interviewees were only analysed by the map and the interviewee
was not given the chance to freely elaborate his or her understanding of this non-knowledge. An
appropriate question might have been “what do you know about this non-knowledge”. Hence the
work about non-knowledge remained rather closed, as only few new thoughts were allowed in. This
perceived failure in the questionnaire could however be smoothened by the results from the second
exploration of non-knowledge (annex VI0), where the question was asked but answers and thoughts
remained rather one-dimensional.
The question for the type of non-knowledge seems to be a case where it was tried to make
hypotheses too early.
The post-normal science diagram was too simplistic to deliver what it was expected to. These
categories could not be cross-checked in relation to the mapped non-knowledge as only one aspect
was mapped and hence the jump to the environment, stakes and idiosyncrasy variables was too big.
The categories were however functional to coincide with interviewees answers. Strong results were
only generated with the time and social question. This was dependent on the fact that most
decisions had influence for more than a year and were taken in a stable social environment. This
however might be related to the content of the question and might not question it´s categories. The
mind-set question´s change in options has been discussed in depth in 3.3.6.
Asking for numerical values in the decision stakes questions had proved difficult for the interviewees
to answer as many decision points are parts of a complex process and it is difficult to distinguish
these as some interviewees hesitated to give numbers for the categories. It is also the case that
human´s system 1, which is in charge of quick categorisation, is blind for statistical values
(Kahneman, 2011). Human-inbuilt heuristics are effective and quick in categorizing (Gigerenzer, Todd
and The ABC Research Group, 1999). However, attaching numerical values had helped to make the
decisions more comparable across interviewees and institutions. A question on permanence of the
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
76
decision or an assessment of criticality (scope, severity, irreversibility), as done in MARISCO, might be
functional for evaluation.
For the volatility questions on the topic, some examples might have been helpful to illustrate rather
stable, developing and abruptly changing topics. This, however, was not done as black swans can
appear everywhere and the question aimed at testing the interviewee´s attitude towards unexpected
change than asking about the topical volatility. It could be advised, if further research is to be carried
out, that the interviewer fits some topics into these three categories to test his or her own attitude
towards unexpected change so that some discussion might develop over the interviewees answer. It
might also help to distinguish between how volatile the topic has been in the past and how volatile it
might be in the future (as done e.g. in MARISCO (Ibisch and Hobson, 2014)), as (Taleb, 2008)
discusses: “Events that are nonrepeatable are ignored before their occurrence, and overestimated
after (for a while).” and “It is why we do not see Black Swans: we worry about those that happened,
not those that may happen but did not.”
The value and mind-set question were also designed to get interviewees self-evaluations, they will
hence be biased. It might be possible with more sophisticated psychological methods to assess the
mind-set more objectively. This would be advised to happen in addition to a self-evaluation. The
interview difficulty question was challenging as sometimes 1 and 3 were very close, when an
interviewee was, e.g. very critical about the offered options, but demonstrated an active
understanding of non-knowledge.
The evaluative questions were functional closing questions and gave the interviewees a chance to
relate everything they had answered back to the initial decision or feel comfortable with their
decision even though they had reanalysed it from a different perspective. The last question, which
was again provocative, worked and made many people laugh, express a desire to formulate it more
humble and think about the limits of their abilities. Abilities can be meaningless if chance comes into
a decision process. However, a decision can also go as planned and thereby depend on the abilities.
In retrospect, it can be, and often is, evaluated if a decision was determined by ability or chance.
Trying to separate the role of ability and chance for not-yet taken decisions would require a better
understanding of the unknown.
Results and Hypotheses
The questionnaire works. It could be applied in future studies. Some hypotheses might need
adaptation and a look into psychological or other guidelines for interview design and decision
assessment might enrich it. The following lists summarises key points of the discussion:
Inbuilt uncertainties occasionally unsettle interviewees
Too many variables assessed for the sample size
Low decision awareness hampers interview flow
Evidence question should be more specific to decisions
Reduction of a decision process to a point is necessary in this method
Difficult to expose heuristics in working life
No correlation between heuristics and decision time, alternatives and why the decision
was taken were found
Unknowns can be inventoried
The map was too prescriptive and impeded open thinking about unknowns
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
77
Type of non-knowledge appeared to have forced making hypotheses too early
Analytical gap between a decision and just one unknown was too large
Post-normal science diagram questions were functional to interviewees
Permanence and criticality of the decision might be aspects of the decision stakes
Examples of topical volatility might be helpful
Topical volatility should be distinguished in past and future changes
Different methods might allow a less subjective assessment of mind-set and values
How could ability and chance be assessed for the future?
4.2.1.3
Interviews
Observed biases appeared to be hindsight, availability and anchoring (“If interviewees struggled with
the question, they were offered to answer if it was rather an intuitive or a rather a rational decision.“
this might have hooked them up on an anchor.) but there were surely others. These are, however,
not considered important as the main aim of the interviews was to gain additional and external
decisions and examples of and conversations about unknowns. In-depth dialogues (3.3.6) about non-
knowledge which generated insights beyond the questionnaire happened in five cases.
Slow thinking was activated before the interview when the interviewees were asked to prepare two
concrete decisions they took at work. The remainder of answers to the question was probably more
influenced by fast thinking (and checked or endorsed by slow thinking), as the interview questions
requested a quick response (put the interviewee in a stress situation).
It was very hard for interviewees to keep focus, usually on the defined unknown but sometimes even
about the defined decision point. This suggests that the human mind (interviewee´s minds) is not
naturally (system 1) aware of specific non-knowledge and specific decision points. System 1 is rather
used to more general and heuristic treatment of decisions and non-knowledge (for which it appears
to be well-suited). Can the transfer of dealing with decisions and non-knowledge into system 2 yield
some benefits?
4.2.1.4
Scheduling and Interviewees
Scheduling interviews with GCM alumni was easy, quick and straightforward in communication and
hence very suitable for the short time horizon of a master thesis. The group was heterogeneous and
somehow sensitised and interested in the topic. A major perceived constraint is the small number of
potential interviewees, so that some interviewees expressed concern that they would be
recognisable from their decisions and answers “Cause you gonna publish this, it might become very
clear from the context that it was me”. Hence a lot of information had to be made unrecognisable
and non-personalised. Therefore it could not be linked to other information and an in-depth analysis
of interesting decisions was very difficult.
If further investigation on the topic might be carried out in a similar way, it is suggested to include
more senior decision makers and different groups. It might also be interesting to ask about life
decisions (as compared to work decisions) in the same group or among different groups, e.g.
children.
4.2.1.5
Clustering and Coding
The coding of information proved difficult in some cases. The goal of functional categories was
reached in most cases; they are, however, not exclusive. The coding was necessary for the PCA but
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
78
binary (final or preliminary), ordinal (years of work experience) and nominal categories (type of
employer) were mixed. The nominal categories did not provide clear hypotheses in the PCA.
Transferring those into several binary categories might yield additional results.
The wording was sometimes difficult as English is not the author´s mother tongue and little
knowledge was acquired about categories generally used in psychological investigation and analysis
of decisions. Words frequently used such as ‘choice’ or ‘judgment’ were difficult to apply. For the
non-knowledge part the need of taxonomy for a new (scientific) field became apparent. Hence the
wording is not yet consistent and defined to not be prescriptive and to not close down or limit the
debate about functional terms at such an early stage.
4.2.1.6
Quantitative Analysis
The lack of background knowledge in classical decision theory, functional categories and interviewing
techniques has generated some “noise” that could not be properly evaluated and comprises
redundant and non-functional information. This noise has, however, to be accepted in a non-
knowledge based approach and is also econical (Hobson and Ibisch, 2012).
A more systematic or a complete analysis of results might have generated additional results but with
a considerably larger investment of time which was beyond the scope of this thesis. The tinkering
with data sets that appeared to make sense and methods that were suitable (available, understood,
quick to use) was functional to generate some results and insights. Also, it has to be kept in mind that
the choice of interviewees was severely limited and that the questions on decisions had an open
wording. This allowed very different datasets to be created that did not have much in common – and
this surfaced in wide scatters in the statistical analysis.
The hypotheses generated by the Principal Component Analysis have to be interpreted with care,
especially when dealing with few and very diverse data sets. The three authors who were mainly
consulted in this thesis, Kahneman, Gigerenzer and Taleb (Kahneman, 2011; Librevis, 2006; Taleb,
2004), insist on and publish about statistics. Statistics should be applied with care and bias-
consciously. The human mind (especially system 1) is not made for inferring from statistics. Hence it
is easy to say that the Principal Component Analysis objectively analyses the data and looks for
regularities but as soon as humans look at those hypotheses, which cannot be prevented, bias
occurs. There are too many biplots and versions to analyse them all, the samples are too small
(especially in groups), the input data have limitations (e.g. with nominal, binary and ordinal
categories mixed) and random patterns might emerge. A careful analysis with “value” and “thought”
applied was carried out and the results cautiously described. With more input data, a Principal
Component Analysis is suggested, as it is able to see more in the data than humans can because it is
purely objective (or purely biased in the way it is programmed and not unsystematically subjective).
In the statistical analysis most variables caused little more than what seems to be “noise”. Whether
the categories (possibly with a revision) are really random could only be demonstrated with a larger
sample.
The bar charts of the non-knowledge map were limited and inefficient when they should be set in
relation to other indicators and groups. Accordingly only one graph is displayed.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
79
4.2.1.7
Limitations
A known approach had objectivised the design of methods and hence reduced the self-referential
tendencies. This was not possible for the non-knowledge map and was tried by incorporating
literature on heuristics and decision making. Here, difficulties arose as disciplinary psychological skills
were absent and the question required an open assessment of decisions and heuristics.
Non-knowledge based approaches should not be the rule, but they can yield additional results from a
new perspective. As the rat in the maze (Figure 5) and many other results from research on heuristics
show: counter-intuitive (or counter-rational) behaviour can yield benefits if applied by a few.
The non-knowledge based approach proved functional to explore non-knowledge. For heuristics and
general decisions some limitations became clear that had possibly been easier to solve by using the
knowledge (e.g. categories) already generated by science. However, the categories suggested here
(3.3.8.2) might be more functional for relating decisions to non-knowledge. Furthermore, the open
questions facilitated the dialogue.
The chosen methods could not deliver worthy results on heuristics (and biases) as these would need
more sophisticated, in-depth and psychological (disciplinary) research (a knowledge based
approach). Further limitations lie in the design of the methods (e.g. the target group). The open
questions also lead to similar answers and few examples were found where an interviewee had not
acquired knowables (which is often understood by laypeople and scientists as a common form of
non-knowledge).
It became also clear during the interviews that group decision making would need different tools of
analysis as the process, point and actors are a lot more complex. Especially if the actors are
institutions or larger entities, single decisions are so decentralised and cannot be viewed in their
entirety. The concerned interviewee furthermore reflected that the actors are highly predictable and
everything would be super-institutionalised; but only because the actions can be anticipated, it is not
clear why something comes out or not – complexity at its best. With a large number of decision
points, tipping points emerge that sometimes cannot be foreseen but often the outcome of group
processes is clear as well, but not the process to get there. The process is often influenced by power,
coincidence and many knowns and unknowns. Big and complex decisions with group level influence
took place in a very stable environment and had low decision stakes.
Finally, it has to be kept in mind that science provides just one form of knowledge. It is limited in its
methods and its perspective, even though it embraces practical, traditional and other knowledge.
Berry (2008) warns that science cannot correct itself as quickly as sometimes needed. Taleb (2008)
states “a thousand days cannot prove you right, but one day can prove you to be wrong (…) it
remains the case that you know what is wrong with a lot more confidence than you know what is
right”. It was attempted to maintain a critical scientific perspective, what Ibisch and Hobson (2012a)
call “systemistic precautionary pessimism” and what Taleb (2012) calls sceptical science with
subtractive empiricism.
Can a convenience sample with convenience literature really deliver inconvenient results?
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
80
4.2.1.8
Recommendations for a Further Use of the Methodology
Generally, the used methods are recommended for further application in exploring this topic. An
adaptive use of the methods is recommended likewise.
A larger sample of interviewees (e.g. 1000) should comprise various subgroups (e.g. children,
investment bankers, scientists working on robotics, philosophers, psychologists, different
nationalities, junior and senior decision makers, managers…). It might also analyse work and life
decisions. Such a larger study would need disciplinary (psychological) and methodological knowledge
input without losing the transdisciplinary and open non-knowledge based approach on other aspects.
A backward approach might also be adopted with more decision-aware decision makers (probably
more senior decision makers or decision scientists) that could be asked for specific non-knowledge
types or forms.
Experts or specialised scientists might also be interesting interviewees to find out unknowables and
soon-knowables as they are used to pushing the knowable and work on the edge of knowledge.
4.2.2
Results
In the results section, exemplary decisions and interviews, as well as answers to questions from the
questionnaire are discussed.
The methods proved useful to explore the non-knowledge map and could give insights on the
exploration of non-knowledge literacy. The gap between the decision and heuristics to the mapped
unknowns (which always remained just one among many unknown aspects) was larger than
expected, so that the heuristics and environment connections could not be strongly observed.
The general approach could be criticised, as e.g. Sutherland et al. (2006) state that uncertainties in
policy-making (which is decision making) should not be solved by science. Baron (2008) states that
good thinking would be a question of philosophy and design – not of science. It is argued here that
not the decisions should be based on science but science can observe patterns and collect
information for meta-issues.
4.2.2.1
Interviews
In the analysed indicators no dangerous patterns could be observed that deserve in depth discussion.
That might depend on the small sample, lack of a comparison group and open quest for any decision
situation at work. If decision situations were more comparable some patterns might emerge with the
analysed data. The main focus of this thesis is the work with the non-knowledge map and the
exploration of non-knowledge literacy. Furthermore, there are not enough data to comprehensively
draw conclusions about the relations between assessed variables and unknowns (as discussed
above). Hence, the discussion of interviews is skimped.
As no assigned variables (decision taken alone, supported or in a group, final or preliminary decision,
complementary, novel etc.) surfaced in the PCA as suitable components, speculations about these
are abandoned. To most individuals they seem reasonable and possible factors but if they are really
important indicators could only be verified by a larger sample. Some ideas linked to the literature
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
81
• Classical distinctions and groups could be set up from general information, such as attributes
of older and more experienced interviewees or differences between men and women
(McKinsey&Company, 2015).
• Rationality is not a common term in psychology but is commonly implied (Baron, 2008), it
was however used here, even though only as one part of the system 2 influenced decisions.
Intuition is more flexible than perception (Gigerenzer, 2007a). Using intuition can hence be a
good choice, even though perceived information point into a different direction.
Interviewees describe the (trained) intuition that is able to deal with non-knowledge. The
experts often seemed to be too deep in the topic to analyse a specific decision point and let
go of the complexity (complexity trap).
• Interviewees were not given the chance to complain about short time or relate limited time
to them having non-knowledge (which happened only in one case). However, the PCA did
not show any connection between limited time and non-knowledge indicators (e.g.
knowability). This might be worth a more in depth analysis if a larger target group is used in
later studies.
• Many decisions were one-time decisions that had only limited feedback as (Gigerenzer,
2007a) suggests. How they can be dealt with might form a specific line of enquiry.
No difference could be observed between interviewees who said they based their decisions at work
on evidence and those that said ‘sometimes’ or ‘no’. It is however remarkable to see that half of the
interviewees had an absolute opinion (yes or no). Scientific evidence is a complex concept that seems
to be rarely discussed by practitioners and would need explications and reflections according to
concrete examples, or be broken down into different aspects such as the reason why to make
decisions and the decisions itself. “Remember that the burden of proof lies on someone disturbing a
complex system, not on the person protecting the status quo.” (Taleb, 2008) The burden of proof,
the burden of evidence: using evidence to decide is like evaluating heuristics against “objectivity”. It
might just be the wrong framework. Evidence is reductionist by definition.
Results in science are often late and usually preliminary. Evidence and policy making both embrace
ambiguity. Other knowledge forms (that are less self-critical) do often provide more useful advice for
action than evidence. Most interviewees appeared to know this but some might get blinded by their
knowledge if asked directly.
The analysis demonstrated that we do not have to go far to find complex systems-related
uncertainty (as asked for e.g. by Ibisch, Vega and Herrmann (2010)). It is assumed that evidence-
based knowledge had been more difficult to find, even when asked concerning the knowns (“What
did you know when you took the decision”).
The low decision awareness in several interviewees was puzzling. It could be discussed in a way that
many decisions do not really matter – it only matters that something is decided. Those might often
be unconscious or daily decisions and they might often be preliminary. There might be many
decisions taken “just like a stroll” with little effort (Kahneman, 2011). And why did several
interviewees express regret or embarrassment about the chosen decisions?
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
82
The low awareness of global-change relation suggests a gap between the working realities and the
study program. It could be explained by larger systemic interactions and connections - a complex
global system, the management of which is strongly decentralised. Or as Ibisch, Vega and Herrmann
(2010) put it: “Possibly, it is not realistic to expect substantial changes to occur as a top-down
process. Paradoxically, global environmental governance will (also) have to start in the form of
multiple bottom-up processes. ” Some interviewees could accept or find a global change relation of
their work, but they were not asked more in depth. The simple assumption was used that global
change managers that work in their profession, in the environment sector, take decisions that are
somehow related to global change.
The level of abstraction between the decision process and the defined decision point proved
annoying at times but was a necessary retention of complexity, a continuous search between chaos
(process) and order (point).
No further classification seemed functional or feasible for decision conditions without additional
research into categories which are suggested by experts. Hence, the decision conditions help the
reader understand the decisions but have no sensible numerical value.
Decisions taken by Global Change Managers were usually not directed and were often not between
fixed alternatives. This suggests Global Change Managers have to take complex decisions. The large
number of decisions where alternatives could be influenced is a good precondition for flexible
decision making. According to Dean (2008), when alternatives can be influenced, it depends on the
boundaries which the individual sets or perceives. No alternatives are difficult to deal with whereas
too much information leads to too many alternatives and excessive confusion. Ibisch and Hobson
(2012a) also state the human inability to deal with all the alternatives created by the knowledge
explosion.
How interviewees took their decisions, what was known and what was unknown appeared to be
unproblematic.
Post-normal science accepts cognitive limitations (e.g. in perception and perspective) as a key
concept (Ibisch, Vega and Herrmann, 2010). The quest for decision stakes has probably been
influenced by different perspectives so that outcomes are not necessarily comparable. Systems
thinking, which Global Change Managers apply to varying extents, dissipates boundaries between
formerly well-defined spheres of influence.
The assessment of “love, will and passion” for decision investment was not done as these are
difficult to measure and usually not limited or linear as time and money are. Much of it might also be
covered in the idiosyncrasy questions.
It was perceived as surprising that, on the social volatility, no interviewee complained about their
boss´ time and availability.
Ibisch and Hobson (2014) discuss the need for a change in mind-set. This change in mind-set would
be different in strengths between cultures. For dealing competently with the unknown, it would be
oriented towards the future. This would allow more flexible strategies. Moral diversity is the rule.
There are knowledge and ignorance-based ethics, morals and religions (Peterson, 2008). They
provide values and often a mind-set to individuals. Rule-based ethics can facilitate dealing with
moral non-knowledge, as discussed by Peterson (2008): “A rule-based ethic might assert that the
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
83
torture of prisoners is always wrong, even if it may generate vital information, or that telling the
truth is always right, even if it will hurt someone’s feelings. Nor do rule-based ethics require
knowledge about the details of the situation since, once the right principle is found, it should be
generally applicable. Being generally applicable, in fact, is what defines it as “moral.””. Peterson
(2008) names Buddhism as an ignorance-based ethic: “Right action can and often should precede
right knowledge. (…)There is no need to wait for certain knowledge to begin action, and, in fact,
knowledge is generally not possible apart from practice. (…)Understanding can follow action,
(…)moral judgment need not wait for full knowledge”. It could be argued that aspects of non-
knowledge literacy can be deduced from ignorance-based ethics without adopting a complete mind-
set.
Many interviewees explained flexibility and adaptability as high values. By using their mind-sets,
individuals and groups develop routines of thinking, analysing and acting (Geiger, Kreft and Ibisch,
2012). Routines finally are manifested patterns that emerged out of (un)directed behaviour over
time. They are highly functional for human survival but challenging them is equally important as
following them. Such a regular questioning of and escaping from routines would be part of an open
mind-set. Global Change Management attempts to teach an open mind-set. However, mind-sets are
influenced by many factors beyond the study programmes reach and are something very personal. It
might hence be a characteristic trait of the target group that their mind-set and values were
explicitly said to have influenced their decisions. This is ambiguous. However, it might be a success of
the study programme that interviewees were able to perceive, admit and often be proud of the
influence of their mind-set and values on their decisions.
Taleb (2008) accuses the tendency to underestimate the role of luck in life. It has also been
investigated that humans tend to attribute successful decisions to their abilities, whereas
unsuccessful decisions are attributed to luck: “For example, a surgeon decided to do an operation,
with a known probability of success, and the operation either succeeded or not. Subjects judged the
surgeon to be a better decision maker when the operation succeeded. The success, of course, was a
matter of luck. The surgeon had taken a “calculated risk.”” (Baron, 2008) Many interviewees that
rather attributed the good results of their decisions to their ability might have fallen into this trap. In
real life, luck and ability cannot be objectively separated as they include too many unknowns
(especially counterfactual unknowns “what had happened if …”). Being humble would be one way to
navigate between abilities and luck. Humility is frequently discussed (e.g. by various authors in
(2008b)) as a way of dealing with the uncertain, and Berry (2008) describes it as a chance to change
oneself and a way to confront one´s ignorance - a place where more authentic knowledge, such as
experience or traditions, applies.
4.2.2.2
Exemplary Interview
The interviewee seems to have a strong mind-set shared with the institution, which can be observed
throughout the whole interview. Several “soft” methods, especially the involvement of peers, are
used. It took a long way to develop the “what will change” non-knowledge out of several unknown
possibilities and eventualities (e.g. someone could have left the team).
The retrospective analysis in the interview deepened the understanding of the decision. It also helps
understanding non-knowledge about the future in general. “I do not know what will change” applies
generically. Everyone can of course make assumptions about future changes, but it will remain
unknown until the time has come. The interviewee´s lax “I don’t care” – it is not relevant what will
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
84
change – appears to imply a high form of flexibility and adaptive management. It seems to fit with
the decision being preliminary. The interviewee did not plan to change the decision, but considered
to adapt it if necessary. Perceiving every possible change as potentially benefiting suggests a
particular mind-set of the interviewee and might have been very different for other individuals. A
static method (and or mind-set) might have judged the non-knowledge relevant with a threat
potential.
The conviction that people across a wide (global) community would be influenced by the decision
might stem from the mind-set and systemic thinking of the interviewee. Influence was defined in a
network way, as not only the influence generated from the coached team but also the interviewee´s
institution and the involved peers were considered. The interviewee was very clear that changes in
the institution happen at a daily basis but that the institution would fit in the first category in terms
of size and age. The changes are, however, developments rather than a volatility as it could be
observed in smaller institutions. The “extremely structured” institution provides guidance and safety
to the interviewee even though it is “highly adaptable”.
Is a 100% evidence base for decisions at work possible? Is a strong institutional and value-driven
mind-set beneficial?
Figure 17 shows the graphical representation of the unknown “I do not know what will change”. It maps knowledge and
non-knowledge as part of one continuum against time. The perspective of the interviewee is blue, a more general
societal perspective is black and the handling is green. The decision point, the interview day and the moment of
manifestation are indicated. Manifestation is when the unknown turns into something known, it is the moment of
revelation. For this unknown, manifestation is the moment for which the unknown is relevant. Before the interview, the
personal recognition of the unknown was limited (dashed line). This generic unknown was recognised in the interview. It
is assumed to remain recognised until the moment of the coaching (manifestation). Simultaneously, the change might
(dashed) already have happened or be known by others (existence or societal recognition). For this unknown, ambiguity
overlays all lines. The interviewee perceived the unknown change as potentially benefiting. The proposed handling was
scenario thinking, adaptive management and trust in own capacities. It bridges the gap to the point of relevance.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
85
Figure 17 exemplifies the graphical representation of the unknown “I do not know what will change”.
Principally, it could be applied for other unknowns. All initial dimensions can be found. Temporality is
in the decision point related to the relevance, manifestation. Ambiguity is in all straight lines and is
displayed by curves. Knowability is the distance between personal recognition and existence or
societal recognition (which might be absent). Solution relevance lies in the manifestation and the
point in time where the unknown becomes relevant. Threat potential and intentionality are a
function of handling. Recognition is the primary indicator and the social and geographical distribution
is in the existence or societal recognition. Handling bridges the gap between decision point or the
interview day and the manifestation (often a gap in time but also in intentionality or availability).
This graphical representation was born out of the second exploration of non-knowledge and the non-
knowledge map. For this exploration, “random” questions and tasks that were aroused at the end of
the work on the thesis, were answered and reflected upon. Those were carried out or formulated
and answered in annex VII.4.
4.2.2.3
Exemplary Decisions
The particularly interesting or illustrative decisions or their combination with unknowns are
discussed in this section.
A. A surprise, even in work context, is a common mean and an example of intentional non-
knowledge. Surprises are usually potentially benefiting.
B. The fact that one interviewee broached the issue of irrelevant non-knowledge for decisions
taken according to a directive opens up the discussion on the stability-generating function of
directives. Creating directives according to which decisions have to be taken are tools that
reduce volatility, they are for example applied when students are selected or evaluated
according to defined criteria. Applying directives (a type of heuristic) is considered easier
than deciding in open (we really do not know how to do it) or semi-closed environments (you
have to take the decisions but you are limited by your environment, e.g. a boss, a budget
etc.). These kinds of directives (or heuristics, simple rules) can limit creativity and favour one-
size-fits-all solutions. Eisenhardt (2015b) and her colleagues promote simple rules (heuristics)
specifically to navigate complexity, maintain flexibility and facilitate decisions.
C. The blindspot about environmental benefits retrospectively recognised by one interviewee
appears to demonstrate how a change in values and knowledge can impact a person. It might
back the discussed value- and mind-set loading of the Global Change Management study
programme.
D. Decisions guided by limits are also simple heuristics: I did not have more time and capacities
so I just took what was familiar (compare familiarity heuristic 2.1.3).
E. Stakeholder involvement opens up completely new information sources and new non-
knowledge and often causes ambiguity to surface, as it allows various perspectives. The
decision was one of the few decisions that were backed by methodological (in this case
spatial) analysis.
F. A decision taken based on values and mind-set without evidence - does a humble end justify
the means? A good intention but based on half-knowledge. An outlier in the interview series.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
86
4.3 Mapped Unknowns
The mapped unknowns were rather unimpressive. They did not hamper the decision making. They
did not cause much philosophising. They were intuitively understood.
The observed strong difference between work and real life decisions might not only be grounded in
methodological issues, but in the fact that many biases do not matter in real life, which is less
institutionalised. Giving good reasons for your decisions is crucial at work; in private life it is often
requested but not necessary. Rieskamp and Hoffrage (1999) discuss the influence of social context on
decisions in terms of importance and accountability – the need to justify one´s view. They found that
the need for justification would often lead to increased effort with the aim to gain surplus accuracy.
At work, people are accountable for their actions. Many decisions might have more impact and be
considered more important than personal decisions. Accountability reduces complexity. So does an
institutionalised framework (an employer), e.g. by providing structures, stability and social control.
Berry (2008) proposes a disembodied corporate mind, which e.g. an institution or any combination of
two or more minds has, and suggests that this is in many cases a lot better than any singular mind.
However, Berry (2008) also discusses that it could be much worse, as it can cause more (unsolvable)
problems, is narrowed by profit or power and is “arrogantly ignorant by definition”. It is not humble.
Dean (2008) discusses that an institution holds less knowledge than the sum of its parts, as the
institution cannot know everything every individual knows; emergent properties might still
materialize out of an institution. Taleb (2012) says that it is a lot easier to fool a multitude than a
single person and Gigerenzer (2006) discusses the influence of institutions on individuals. This
reduction of complexity might be practically observable (as suggested by Rieskamp and Hoffrage
(1999)) or just in conversation (I justify my decision with system 2 but finally it was based on system
1 – I wanted it this way). In private life communication, wellbeing and a good treatment of others are
frequent drivers for decisions. At work, these are rather subliminal. In work decisions, principal
drivers might include accountability and pressure from hierarchy, civil society or an institution´s
purpose or power.
It became clear that there are qualitative differences in the mapped unknowns. The exemplary
interview (compare 3.3.5 and 4.2.2.2) provides a prototype of generic non-knowledge. Other generic
unknowns are I don’t know if I have a blindspot in this issue (e.g. an unrecognised mistake) or any
counterfactual I don’t know what had happened if x. Once recognised, understood and mapped it
might feature current and future decisions on any topic. Other unknowns are semi-generic, as they
depend on a mind-set, e.g. if there is a God or a higher force, if there is a beginning etc. Most
mapped unknowns were specific to the situation but occurred repeatedly: I don’t know the
circumstances (someone´s intention, how a situation was, is or will be etc.), I don’t know the
consequences of my decision (how will the system and it´s elements (e.g. someone else) react to my
decision, this includes also non-knowledge if I can reach my goals) and I lack complementary
knowledge (I don’t know how to do it, if to do it or what exactly to do (I don’t know how to decide)).
The categories ‘future’, ‘others’, ‘information’ and ‘value’ did not proof useful and were hence
dismissed (any further discussion about them would anchor and hence bias the reader). Peterson
(2008) asks moral questions that describe the dismissed categories ‘value’ and ‘others’ –moral non-
knowledge might be a useful category and an accepted term.
The suggested categories for (non-)knowledge, complementary, consequences and circumstances,
are functional related to the analysed decisions. A how to decision is complex and needs
complementary (non-)knowledge as well as (non-)knowledge about consequences and
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
87
circumstances. If to and what exactly decisions only require consequence and circumstance (non-
)knowledge as the action for which complementary (non-)knowledge would be needed is clearly
defined.
Unknown circumstances were specified in more than half of the decision situations. This suggests
that interviewees could intuitively deal better with unknown consequences and a lack of
complementary knowledge and hence perceived them as less relevant. It might be worth some in-
depth investigation if unknown circumstances are perceived more problematic in decisions.
Interviewees seem to be used to rather stable environments (compare 4.2.1.2) and dealing with
stable circumstances is psychologically easier for most people. Taleb (2012) challenges this by
proposing that taking options that are psychologically uncomfortable but might have a sense of
adventure and thrill are necessary at times. It can however be questioned if a focus should be laid on
making circumstances safer, as this reduces volatility and (might) increase the risk of black swans and
bigger collapses.
Most unknowns were bounded, they could be somehow approximated from experience (e.g. how
someone will react or what is a better option).
Berry´s (2008) ignorance classes (compare 2.1.5) are partly covered in this thesis. Fearful ignorance
(and the one caused by laziness) relates to intentionality. Inherent ignorance would be the one
caused by mind-set and values, which are asked but not assessed. False confidence would lead to
mistakes. Together with half-knowledge and wrong knowledge, they remain outside the scope of this
thesis. The same applies to for-profit and for-power ignorance where the non-knowledge does not
relate to the individual (e.g. decision maker) but is used like a tool.
Berry´s (2008) knowledge categories (compare 2.1.4) were more present in the thesis. What is called
“sitting-duck knowledge” was mapped once (complete range of criteria) and two or three swimming
ducks (practical details about standards) were named. This category is closest to the dismissed
“information” category, but this also included landing ducks (e.g. who will be in government). If it
was not a methodological bias, “sitting-duck knowledge” was either not important or it is not a major
unknown in a knowledge society. The uncertain knowledge from experience and use can be found in
complementary (non-)knowledge. Traditional, as well as religious, knowledge is a hybrid form (is
largely complementary, but can be about circumstances or consequences) that can enhance a mind-
set; it has the discussed limitations for the future. Instinct or inborn knowledge provides
complementary how-to knowledge. Intuitive knowledge relies on inborn or trained heuristics and
was assessed by the question “how did you decide”, it can provide all three knowledge categories.
Conscience, inspiration, sympathy and bodily knowledge are somehow similar and could not
explicitly be grasped within the scope of this thesis, although some interviewees referred to these in
terms of (non-)knowledge. Neither was counterfeit knowledge or plausible falsehood.
The following quotation from Ibisch and Hobson (2012a) appears unclear when it suggests the
combination of a knowledge inventory with a non-knowledge mapping is those are meant to be
equivalents: “The development of a knowledge inventory can be combined with a non-knowledge
mapping exercise that identifies as well as classifies the various types of non-knowledge within a
given context”. Results from the questionnaire suggest that a (non-)knowledge inventory (questions:
What did you know in the decision situation, what did you not know in the decision situation) is a
useful analysis before a specific (un)known is mapped.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
88
The following hypotheses and questions can be derived from the preceding discussion of the
analysed unknowns:
Assessed unknowns are “unimpressive”
There appears to be a strong difference in handling of work and life decisions
An inventory is a useful heuristic analysis before a mapping
Inventories and mapping can be done for knowns and unknowns
Generic unknowns can be mapped
It is suggested that (non-knowledge) can be about circumstances, consequences or it can
be complementary
Are unknown circumstances perceived more problematic in decisions than unknown
consequences or a lack of complementary knowledge?
Should we get more used to less stable circumstances?
In the mapped unknowns, Berry´s (2008) knowledge categories were more present than
the suggested ignorance classes
4.4 Non-knowledge Map
After the interviews, many dimensions remained unclear and insights vague. The discussion was
largely informed by the second exploration of non-knowledge and the non-knowledge map,
presented in annex VII.4. Asking gradients of each dimension separately might have been a one-
dimensional approach. From the beginning it was envisioned to find links between dimensions. These
links are then presented in the reorganisation phase 5.
The seven investigated dimensions of the non-knowledge map are applicable to the analysed non-
knowledge. However, some dimensions are more closely linked than others. The map was developed
to a board game (annex VII.1) and a graph (exemplary in Figure 17), both of which can be used for
tinkering and exploring non-knowledge. A trial of a naturalistic representation of the unknown (an
ocean of unknowns, a river of approximated unknown, an island of defined “sitting-duck knowledge”
etc.) was dismissed as it appeals strongly to system 1 which offers a quick and intuitive
understanding and is easy to overshadow the more critical system 2. Understanding a graphical and
diagrammatic representation appears to have been learnt by system 2, which would hence be
activated. This is why abstract representations are suggested.
The map only works for closely defined unknowns related to specific decision points. This reduction
in complexity at one end is necessary to allow complexity in other parts. For every decision, there are
an infinite number of unknowns and not all can be mapped. What the map provides is an active
quest for system 2 to put effort into understanding a few unknowns, with the hope system 1 learns
about those typical unknowns and can find them quickly and derive management solutions, that can
be made explicit and endorsed by system 2.
Some of the adaptations made for the map´s application to decisions (as used throughout this thesis
and presented in Figure 11) could be changed back to the initial version. The game (annex VII.1) for
example was only in need of the additional gradients for retrospectively recognised and extends
through the present. All others can be explored by questions from the initial compass like map.
It became clear that some dimensions are more static than others. Temporality, recognition,
intentionality and relevance are static, as they strongly relate to the individual or the decision point.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
89
Other people would usually not challenge this. Knowability, ambiguity and threat potential can be
changeable and indefinite among different people, as they tend to depend on an individual´s
perspective.
Finally the map requires the user to make decisions about unknowns. The user does not know much
about the unknown but tries to approximate it; in some cases (with the static dimensions) this is
easier. For ambiguity, knowability and threat potential it often remains a guess.
Only one map was mono-polar. The hypothesis that black-parts of the map are more relevant cannot
be confirmed or disregarded. If the black parts of the map are more relevant to sustainability cannot
be the question in a decision focused enquiry and would hence need a new phrasing “more relevant
to ?”. This would however not be the case for intentionality as it includes a conscious decision for an
unknown. Why the initial map is narrowed in its application to sustainability remains unclear.
Adaptations discussed here would allow using the map for at least any decision related unknown.
Some combinations of the non-knowledge dimensions were more common than others. Front-
runner was eight combinations either potentially benefiting (4) or with a high threat potential (4)
where the other dimensions were: future related, ambiguous, now unknowable, relevant,
unintentional and recognised. More patterns would need larger samples but could then possibly be
linked to decision and types of non-knowledge. Probably a distinction had to be made between what
is now called static and perspective-dependent dimensions.
Intentional and retrospectively recognised non-knowledge were exceptional cases. Intentional
unknowns seem to be perceived negatively in a knowledge society even though the author assumes
them to be an awesome. That so few blindspots were mapped might depend on the nature of the
question “what did you not know in the decision situation” which sets people into thinking of that
specific situation and associated recognised unknowns. Probably there are also usually more
recognised than retrospectively recognised unknowns in a situation, as retrospective recognition
requires some sort of posterior thinking about the decision which can be caused by external events
or internal thoughts but it can be assumed that most people do not routinely systematically
contemplate taken decisions. Blindspotting would be a tool for this (Ibisch, Geiger and Cybulla, 2012).
The analysed decisions from the interviews were past decisions that could have retrospectively
recognised unknowns. It was successfully applied as a decision aid to current decisions (compare
annex VII.4).
The next section discusses the dimensions separately. The following hypotheses can be derived from
the preceding discussion of the map:
The seven investigated dimensions of the map work for specific unknowns
Abstract representations should be preferred
Knowability, ambiguity and threat potential seem to differ according to the user of the
map
The map could be applied to unknowns related to present decisions
A larger sample of maps might allow finding patterns in distribution of dimensions
In posterior evaluation, blindspotting might help finding unrecognised unknowns
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
90
4.4.1
Temporality
Temporality is inherent and one-directional in human perception. It cannot be avoided or influenced
in relation to a specific unknown. As soon as the unknown is defined it should remain static and is
often used as a straightforward reason for non-knowledge. But assumptions about the future and the
past are possible (-> Knowability).
Temporality and knowability are either very closely linked or they are commonly confused. There is
the temporality of an unknown – a specific point or period in time in which the unknown has
relevance and there is knowability which depends on the temporality (when was, is or will it be
unknown). In the graph like representation of the map (Figure 17 as an example) this distinction can
be made.
Faulty decision making caused by not considering possible future nonlinear system dynamics, as
Geiger, Kreft and Ibisch (2012) suggest, could not be observed but are probably an issue.
• Temporality is one-directional and only the relation to it can be changed
• Temporality and knowability appear to be closely linked
4.4.2
Ambiguity
Ambiguity is a dimension which was very difficult to grasp for many, but the answer “it is ambiguous”
is often clear. This suggests that system 1 knows exactly what ambiguity is but system 2 has its
difficulties with the term and where exactly it refers to: the associated known (I assume that if I
would know what I do not it would be ambiguous), the expected effect (it is ambiguous how this
unknown might affect me) or to the unknown (I am clear about not knowing this – in contrast to
somehow oscillating between known and unknown, approximating something uncertain).
Analysis of 40 unknowns demonstrated that ambiguity can be found in the majority of the cases. The
graphical representation (Figure 17) of the map postulates that all lines might be overlaid by
ambiguity.
Ambiguity can distort the assessment of knowability, e.g. when an unknown is perceived as so
ambiguous that it cannot be imagined that someone else could securely know this.
Acting under ambiguous non-knowledge often leads to deciding for better known problems (Geiger,
Kreft and Ibisch, 2012), this could not be confirmed by the results of this thesis, probably due to
methodological limitations; it probably has a point.
• Ambiguity is difficult to grasp and seems to appear across dimensions
4.4.3
Knowability
In MARISCO (Geiger, Kreft and Ibisch, 2012) distinguish knowability into very well known, somewhat
known, little known, not known/not knowable. This is distinct from the categories proposed here for
decision making. It integrates ambiguity. For decision making it was argued, it does not matter much
if something is well known or somewhat known if I do not know it in the concrete decision situation.
It was also attempted to assess knowability separate from ambiguity. This is also reflected in an
updated manual (Ibisch and Hobson, 2014) where the latter two are changed into not known but
theoretically knowable and not knowable. It could be argued that these limitations in the situation
are characteristics of the situation and not of the knowability.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
91
This dimension would probably gain from a better understanding of what knowledge is. Is something
approximated, something defined, something perceived known? Is what someone else claims to
know known? Is it, if that someone else is more experienced with the system or has higher
intellectual capacities? Must knowledge be verifiable by external sources?
Knowability was found to be relative at times. An external approximation, e.g. by the boss, was
sometimes understood as knowable.
Complexity, temporality and ambiguity are frequent reasons for unknowables, whereas knowables
are usually explicitly or implicitly defined (society has reduced complexity), are built on shared mind-
sets, on evidence and scientific findings (“sitting-duck-knowledge”) or on inherent knowledge forms.
• Knowability might be less dependent on individual perspectives if there was a shared
understanding of what knowledge is
4.4.4
Solution Relevance
In terms of solution relevance the category “would be solution relevant” represents an example of
confusion between knowns and unknowns, e.g. the reflection “If it is generally unknowable it even
questions the category of highly relevant. How can something that is fundamentally unknowable be
relevant to the solution or decision? Should I not (and am I not) able to decide without knowing it
(exactly)?” and the associated example “the decision makers of a nuclear phase-out do not know if a
nuclear disaster will strike before their decided phase-out; if they would know the date they could
phase out a day before. However, relevant for the solution are only heuristics and guiding principles,
e.g. the precautionary principle – it is highly disputable if this is really not relevant or if it would be
“relevant (if it was knowable)”.
Something unknown can be highly relevant. And this relevance closely links to a decision process.
Often the most crucial unknowns mapped were evaluated as irrelevant when asked specifically to the
decision point. This appeared to be due to users of the map relating their unknowns to a different
decision point within the larger process.
Often decision conditions or the framing of the decision can be in a way that what I do not know is
not relevant. Decisions can be framed in a way that with or without a potentially benefiting non-
knowledge it would be worth taking the risk, even though objectively looked at the situation it only
made sense if it was positive. Hence, it is not only about decision making but also about how to
formulate and perceive the decision. Decision making starts long before.
In later stages of this thesis, the understanding of solution relevance oscillated around the terms
“relevance”, “solution relevance” or “decision relevance”. Relevance might not be enough, as it
provokes the question “relevant for what”. Solution is more generic than decision and solution
allows a bigger picture than decision does. Solution relevance would be more useful to open
situations where discussion starts from an unknown without a predefined decision. This however,
has not been tried for unknowns (it works for knowns, compare annex VII.4).
• Something unknown can be highly relevant
• Relevance of an unknown can be changed for current of future decisions by framing
4.4.5
Threat Potential
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
92
The antifragile design of this dimension requires deciding for one perspective on the issue. Clearly
defined unknowns (no “or” in the unknown) enclose this perspective. Changing perspectives appears
to require a flexible mind-set. Weighing those perspectives can be done by attaching probabilities to
beliefs and assumptions. Many users refuse to reduce complexity so much as to fit ambiguous cases
into either of both categories. Threat potential and potentially benefiting exist simultaneously.
Ideas are often potentially benefiting, the result of more or less conscious tinkering in mind.
Decisions based on ideas are hence more likely to be “convex” (as Taleb (2012) uses the term).
Klein (1999) offers a suitable example "The risks of jettisoning the service module were unknown and
could be catastrophic. Framed in this way, the comparison was between a course of action that was
painful but manageable and a course of action with a risk that was plausible and catastrophic."
Gigerenzer (2007a) suggests that “Imitation can be beneficial in situations with high threat
potential”. However, this “can be” has to be handled with care. While Gigerenzer´s food example
stems from basic human behaviour and environment – it has a high threat potential to eat some
plant you do not know, so rather eat what everyone else is eating. Today´s situations are often more
complex, so that: imitating might either be a bad idea (1), an idea that impedes progress (non-
knowledge as a driver for innovation – I did not know – I just did it – it worked (actually better than
everything else that had been tried before)) (2), a manifestation of black-swan proneness (3) or there
might also be no one to imitate (4).
The threat potential dimension was not perceived as a classical risk dimension. To whom or what the
threat or benefit relates was not commonly questioned. Threat potential only becomes relevant with
handling (the decision to act (and how to act) or not act relating to an unknown).
• Threat potential and potentially benefiting often exist simultaneously
• Threat potential seems to be relevant with handling
4.4.6
Intentionality
As results from the interviews indicate, unintentional non-knowledge seems to be the default. The
question often aroused an initial and reflexive reaction which expressed a lack of comprehension
(“unintentional unknown? häh?”). Once understood, most people appear to be quick in assigning
unintentionality. This would make it a rather static dimension.
In the graphical map (Figure 17) only intentional non-knowledge would be represented as a way of
handling the unknown.
Dean (2008) illustrates intentionality in a context of setting boundaries of consideration: a person in
a boat on the ocean in a storm is able to focus on the most important tasks for survival. On the
discussed example those boundaries are not deliberately set but rather an innate (heuristic) function
of the brain in a stress situation. It elucidates the ability of the brain to (intentionally) ignore a lot to
allow focusing resources on urgent tasks. Humans are usually capable to change their boundaries of
consideration.
Neglected (non-)knowledge can be understood as something within the vicinity of more or less
deliberately set boundaries of consideration. A few rather responsible, self-critical and decision
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
93
aware users of the non-knowledge map seemed to see those boundaries of consideration including
their intentionality as more ambiguous.
• Unintentional non-knowledge seems to be the default and would not need specification
• Intentional non-knowledge would be a form of handling
• Neglected (non-)knowledge, as a gradient of intentionality, might depend on the individual´s
setting of boundaries of consideration
4.4.7
Recognition
Recognition is a straightforward category with the option to have recognised the unknown in the
decision situation or in retrospect. Unrecognised mistakes are typical blindspots. Blindspots could be
called generic unknowns which factor into every decision or action.
Mapping this generic blindspot unknown is one option, the other would be active blindspotting, as
e.g. Geiger, Kreft and Ibisch (2012) describe perspective change or Taleb (2012) tinkering as
blindspotting tools. Geiger, Kreft and Ibisch (2012) show that the boundaries (of consideration) one
decides to draw can create blindspots.
Having forgotten something appears to be located between intentionality and recognition. In the
understanding used here it was, as neglected (non-)knowledge, within intentionality.
Blindspots present an anthropocentric concept of the discoverable unknowns in the entirety of the
unknown. The term blindspots might, however, be misleading as it proposes unrecognised non-
knowledge to be a spot, not a space. But if unknowns have seven to nine dimension, as understood
for now, neither a two-dimensional spot nor a three-dimensional space offers a sufficient concept. In
that sense the sentence “together with the help of science, the number of blindspots diminishes”
(Ibisch and Hobson, 2012a) is questionable. If it is assumed that non-knowledge be endless, as
counterfactual reasoning or scenario thinking suggests blindspots will appear forever.
• The recognition of blindspots seems to be limited by boundaries of consideration
4.5 Non-knowledge Literacy
Non-knowledge literacy does neither promote good thinking nor good decision making, it rather
postulates to draw heuristic principles of dealing with the unknown together and make them
teachable.
It cannot be assumed that someone who could easily decide in an uncertain decision was non-
knowledge literate (this might depend on luck, short-sightedness etc.) and it can neither be assumed
that a very good decision under uncertainty was made by a non-knowledge literate individual. Per
definition, decisions under uncertainty cannot be “perfect” and hence cannot be deduced from the
result. Hence the chosen way to identify elements of non-knowledge literacy were extracts from
literature and the analysed decisions, as non-knowledge literacy does not (merely) depend on good
results (that cannot be known) but should lead to a less confident but more happy way of deciding in
uncertain situations. It should also provide a way that facilitates communication and interaction
among individuals and groups. It is helpful in a work context to explain advantages of the postulates
underlying a decision. Non-knowledge literacy must allow flexibility for action if future changes occur
while maintaining a necessary reliability – as humans (psychologically) need some degree of stability
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
94
(even though it could be argued today´s western world is too stable, compare 4.2.1.2). Non-
knowledge literacy should aim at striving between chaos and order.
As no decision maker in the interviews bemoaned or explicitly spelled out how hard the decision was
according to the uncertainties and unknowns, it can be assumed that Global Change Managers are –
to a certain degree – able to practically (and happily) deal with non-knowledge. In the analysed
situations, they did not use the non-knowledge as an excuse for inaction; finding such cases had
required a different approach. Even though Global Change Managers were, by training, sensitised for
non-knowledge, none of them seemed to have too actively thought about non-knowledge or the
unknown in the situation – they were not applying conscious tools for non-knowledge management.
Their dealing with non-knowledge was rather system 1 guided – and this intuitive dealing with the
unknown is a reasonable starting point to explore the term non-knowledge literacy (compare 2.1.6).
Many interviewees struggled with the questionnaire-inbuilt uncertainties; there might hence be a
need to teach an explicit mind-set towards non-knowledge and uncertainties, as well as their
(potentially stabilising) functions – maybe an ignorance-based worldview. Humility is closely linked to
this. A hint of the interviewee´s humility towards their own abilities in decision making could be
gained from the last question. Further research might explicitly evaluate the humility and non-
knowledge awareness. Counterfactual reasoning trains both.
A change in attitude towards the unknown seemed to be a question of individual mind-sets. Some
interviewees (e.g. the exemplary interview: every unknown change is potentially benefiting) were
positive about the unknown and how it might affect or did affect their decision, whereas others were
rather cautious. This might also depend on their institutions (Gigerenzer, 2006).
How and if interviewees limited their interactions with the unknown could not be assessed with the
chosen methods. It might have been implied at times. However, this protective measure can be
discussed, as it opposes the otherwise promoted proactive dealing with non-knowledge and might
often be impossible. It remains, however, a reasonable tool.
Decisive action was observed in some cases but Global Change Managers seemed to be rather
careful and retained their doubt - at least in mind (the decisiveness of the action could not be
assessed, only what was reflected in the talk). However, the evaluative questions (which might be
biased), pointed to the fact that about half of the interviewees were happy with their decision and
were sure they took a good decision might hint to decisive action.
Non-knowledge itself as a tool was not matter of this thesis but one interviewee lamented about
politicians that use non-knowledge and uncertainties to maintain their position and inaction.
Taleb´s bold ideas were not discussed, but many interviewees might not be in a position to allow
breaking what needs to be broken. They might act in an institutional environment which (socially)
impedes many small mistakes and would react negatively to injected confusion.
From the proposed aspects of non-knowledge literacy in the literature, some seem to be hard-wired,
others seemed to be grounded in a mind-set which most Global Change Managers appear to have
and others are still highly underrepresented and should be taught or tried.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
95
At some points, the analysed principles referred to dimensions of the non-knowledge map (e.g.
recognition and knowability). The map allows a question-based exploration of located or defined
non-knowledge. The proposed heuristic might become a primary hook to teach non-knowledge
literacy.
Taleb (2008) insists “that we got here by accident does not mean that we should continue to take the
same risks. We are mature enough a race to realize this point”; he also warns not to build
frameworks around things that worked by chance.
Geiger, Kreft and Ibisch (2012) describe a situation where non-knowledge hindered action “Some
factors had at first been evaluated as manageable even though the level of knowledge was very low.
Later, this was then re-evaluated because it is unlikely that a factor that is not sufficiently understood
is highly manageable.” could non-knowledge literacy overcome this?
Various authors formulate similar to Perry (2008): “Even so, we notice that countless generations—
billions of people— acting in and through their ignorance over millennia, somehow managed to
survive.” Is this expression of surprise that humans live with ignorance not a deep expression of a
knowledge-based worldview? Is it not that humans need and are inherently able to handle
uncertainties and non-knowledge and have just forgotten or unlearnt that, as the paradox in the
introduction (1, Hobson and Ibisch, 2012) describes.
“My dream is to have a true Epistemocracy—that is, a society robust to expert errors, forecasting
errors, and hubris, one that can be resistant to the incompetence of politicians, regulators,
economists, central bankers, bankers, policy wonks, and epidemiologists. We cannot make
economists more scientific; we cannot make humans more rational (whatever that means); we
cannot make fads disappear.” (Taleb, 2008) – is non-knowledge literacy a way towards
epistemocracy?
A short summary of the findings on non-knowledge literacy is presented in 5.2.3.
5 Reorganisation Phase: Post Analysis Review
There are no known cookbook-style rules for structuring this section as it is not common in natural
scientific enquiry. This section reorganises everything analysed so far into a new concept. This
concept would then be open to modification and further exploitation along the panarchy. As this new
concept would be understood as a result of this thesis, this section is structured along the well-
known lines of methods, results and discussion.
5.1 Methods for Reorganisation
Reorganisation builds on the material generated in the text above. The key material about the
dimensions is presented. Reorganisation is then done by reflections and subsequent definitions.
These definitions are then applied to the 40 unknowns from the interviews and comprise the results.
5.1.1
Material
This section builds on the previous discussion. From the boundaries between knowledge and non-
knowledge it proposes lost and missing dimensions and reorganises and reduces these into a new
concept. The following thoughts and hypotheses are derived from the separate discussion of
dimensions. / Material
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
96
Temporality is one-directional and only the relation to it can be changed
Temporality and knowability appear to be closely linked
Ambiguity is difficult to grasp and seems to appear across dimensions
Knowability might be less dependent on individual perspectives if there was a shared
understanding of what knowledge is
Something unknown can be highly relevant
Relevance of an unknown can be changed for current of future decisions by framing
Threat potential and potentially benefiting often exist simultaneously
Threat potential seems to be relevant with handling
Unintentional non-knowledge seems to be the default and would not need specification
Intentional non-knowledge would be a form of handling
Neglected (non-)knowledge, as a gradient of intentionality, might depend on the
individual´s setting of boundaries of consideration
The recognition of blindspots seems to be limited by boundaries of consideration
5.1.2
Reflections and Definitions
“Knowledge exists as a spectrum between the deeply knowable to the unknown” (Ibisch and Hobson,
2012a). Those spectra are different between individuals and for different situations. It could be
conceptualised that they converge at a level of higher order. At such a level unknowables could be
distinguished from knowables according to their social distribution and geography. For finding such a
convergence, many situation-specific unknowns might need to be clustered.
That knowledge and non-knowledge exist as a continuum and a Principal Component Analysis could
possibly converge the according components into a straight line became clear when the non-
knowledge map was applied to knowns (compare annex VII.4). That the limits of knowledge and non-
knowledge are blurred was also observed when two interviewees said the same thing when asked for
a known and an unknown. Reflections about such hybrid or half-knowledge forms might be needed.
By the graphical representation of the map (like Figure 17) this oscillating or the simultaneous
existence of knowledge and non-knowledge about something can be displayed. This might be the
very aspect of uncertainties.
The graphical representation of the map (Figure 17), has demonstrated how knowability can be
mapped over time. Two points in time became apparent. The first point in time is when an unknown
materialises out of a blindspot. Taking a closer look, the materialisation of the unknown is its
recognition. Such recognition would be perceived in an individual mind first but may also gain
recognition in a collective mind. It might then materialise in form of behaviour, (oral) traditions, text
or other forms of knowledge that endure over time. Recognition is the point in time when a
blindspot materialises into a known unknown.
With recognition, a second point in time becomes visible. This would be the moment when the
unknown would possibly turn into a known. This point would be called manifestation. If the whole
society is considered, it can lie in the past, in the present or in the future. If it lies in the future, the
manifestation remains a guess. For every such guess, a new map could be created (compare 4.4).
In the following, the reflections and definitions that led to the proposed new non-knowledge
mapping heuristic are presented. Figure 18 illustrates the tentatively defined terms:
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
97
Definition: A blindspot is a meta-unknown.
Definition: An unknown is an unknown meta-known.
Definition: A known is a known meta-known.
Definition: Recognition is the (imaginary) point in time when a blindspot materialises into an
unknown.
Definition: Manifestation is the (imaginary) point in time when an unknown or a blindspot
materialises into a known.
Figure 18 builds on the meta-distinction of knowledge and non-knowledge (compare 2.1.5). Meta-unknowns are
blindspots and meta-knowns can either be knowns or unknowns. In the figure, the blindspot is largest as it is
continuously fed by oblivion and the infinite unknown that Ibisch and Hobson (2012b) refer to as dark matter. The
unknown is fed by oblivion of knowns. The (non-)knowledge would be held by an individual, a group or society. The black
processes should be considered for the question at hand. A blindspot can either become an unknown by recognition, or
it can become a known by manifestation. An unknown can become a known by manifestation. Oblivion can reverse
these processes. Blindspots created by oblivion can usually be reversed easily. From now on, terminologically, a
blindspot is distinguished from an unknown.
These definitions are applied over time. The (non-)knowledge of a decision maker has to be
distinguished from the totality of (non-)knowledge held by society.
• Recognition and manifestation are contemplated in relation to time.
• Recognition and manifestation can happen on a personal and a societal level.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
98
This allows specifying the following terms:
Specification: Personal recognition is caused at least by perception. In many cases, it is also
specified by existence and or societal recognition.
Specification: Societal recognition is caused at least by personal recognition of one member of
society. In many cases, it is also specified by existence.
Specification: Personal manifestation is caused at least by perception. In many cases, it is also
specified by existence and or societal manifestation. It can be preceded by personal
or societal recognition.
Specification: Societal manifestation is caused at least by personal manifestation in one member of
society. In many cases, it is also specified by existence.
Specification: Existence needs a medium. Existence can at least be in mind, in space, in text, in
behaviour or in any combination of these. One indicator for existence is perception. It
is assumed that society has a widely shared heuristic understanding of what
existence is.
Specification: Perception can be caused at least by relevance and chance. If a tree falls but nobody
hears (perceives) it, is there a sound? With Heisenberg´s uncertainty principles it goes
into quantum mechanics.
Personal recognition and manifestation are used for decisions. The understanding of existence,
societal recognition and manifestation facilitates the classification of (un)knowns. The following
heuristic distinctions (see Figure 19 for an overview) shall serve to classify (un)knowns on a scale
starting from blindspot and then heuristically set:
Heuristic: Neglected (non-)knowledge would be pure system 1 activity, something like
subconscious intentionality (compare 3.3.1.6); it might just be generated or
observable in retrospect.
Heuristic: Ambiguous (non-)knowledge is open to different perspectives and includes
approximations and assumptions.
Heuristic: Clear (non-)knowledge is neither of the above.
Heuristic: Clearly emerged (non-)knowledge is clear because it builds purely on society´s
concepts that demarcate stabilising knowledge against the unknown, e.g. money,
traffic rules, multiple choice tests and certification.
Earlier terms for ‘emerged’ included ‘defined’ and ‘agreed’. The term ‘agreed’
claims that there was an actor (e.g. society) that has actively agreed that this
now be (non-)knowledge. In a dictatorship this might happen, but
democracy, science and the economy have ways which go beyond ‘agreeing’.
The term ‘defined’ omits the actor but definitions are a strictly defined,
agreed and emerged concept. It is argued here that ‘emerged’ includes
‘agreed’ and ‘defined’ but leaves open how the patterns emerged.
Now, non-knowledge literacy could find concepts of handling in this classified (non-)knowledge.
Handling is understood to be the active human link across recognition and manifestation. It can
accelerate and change the processes of recognition and manifestation. Handling generates
complementary (un)knowns and consequences. Handling is implemented by decisions. Handling
includes non-handling. (Non-)evaluation is the passive human link across recognition and
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
99
manifestation. It contemplates circumstances and consequences of recognition and manifestation.
Evaluation generates complementary (un)knowns and informs handling.
Possibly, there is an unconscious active human link across recognition and manifestation. It might be
system 1 activity which has not passed conscious system 2 endorsement. Temporality and possibly
complexity are non-human links across recognition and manifestation.
The heuristic classification of knowledge and non-knowledge integrates recognition, ambiguity and
aspects of intentionality (neglect, unintentionality) on one scale. These classifications mapped over
time integrate (un)knowability and temporality. Recognition and manifestation integrate the social
distribution and geography of the (non-)knowledge. Handling comprises at least intentional
(un)knowns. Evaluation would integrate possibly ambiguous perspectives of at least potential
relevance, threats and benefits. All nine initial dimensions are considered.
• Mapped over time, recognition, manifestation, handling and evaluation can replace the nine
dimensions of the non-knowledge map for sustainability.
• Recognition, manifestation, handling and evaluation can map (un)knowns of decision
processes with their potential pasts and futures.
Those specifications could be enriched by the following ideas. These ideas would provide starting
points for further research:
• Density would expresses how single (un)knowns are distributed and connected
o Locality could heuristically be distinguished into one, some or many elements.
o Distribution could heuristically be distinguished into concentrated, systematic scatter
and random scatter.
o Interconnectedness could heuristically be distinguished into not connected,
somehow connected and heavily connected.
• Complementarity would express the degree of inexact how-to knowledge useful for the
(un)known. The complementary (un)known would have to be defined and could heuristically
be assessed at least by the following
o Experience: often practiced, some practice, thought experiment
o Ownership: mine, shared, widely shared
o Reliability: proven, certain, uncertain
• Interference would specify how much and if the decision maker could influence the possible
manifestation of the unknown over time. This would require assessing at least the
changeability over time and the capacities of the individual.
• Accessibility would describe how easily available a(n) (un)known is. It could be assessed at
least by density.
• Usability integrates accessibility, and capacities. It is potential handling. It would include
idiosyncrasies.
• Consequentiality would assess the decision stakes.
• Transfer would integrate degradation, decay and elevation
Within the limits of defined, specified and heuristic terms, the assessment would have to allow
maximum subjectivity (as presented in the questionnaire and interviews above). With a large sample,
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
100
this would allow tracking the societal understanding. Once reliable patterns of this societal
understanding emerge, those subjective individual definitions might be transferred into objectively
defined terms. And on it would go.
5.2 Results
This section proposes a new heuristic non-knowledge map for decisions which is written in code. The
exemplary interview unknown “I do not know what will change” is exemplarily coded. A list of all
unknowns from interview decisions and five scenario examples is presented. This is followed by a do-
it-yourself instruction for the coding. The section ends with deductions to non-knowledge literacy.
5.2.1
Exemplary Interview
The proposed new non-knowledge map is built on a very simple heuristic: It maps derivatives of the
initial dimensions against time. This code for decision unknowns maps recognition, manifestation,
handling and evaluation against time. Elements and keys for the code are given in Figure 19.
An imaginary three dimensional space allows mapping the decision´s circumstances (recognition and
manifestation of (un)knowns), the decision´s consequences ((un)knowns over time) and the
complementarity (handling and evaluation of unknowns).
Timing
Meta (Non-)knowledge
X
Decision
B
Blindspot
H
Handling
R
Recognition
N
Neglected
M
Manifestation
A
Ambiguous
E
Evaluation
P
Reference point of X
C
Clear
T
Today
D
Clearly emerged
Figure 19: the tables provide the key to code an unknown. The points in time have to be ordered for the specific
situation. A class of non-knowledge describes R and a class of knowledge M. The (non-)knowledge classes are ordered
according to scale from blindspot (low) to clearly emerged (high). Handling and evaluation are not specified in detail.
Recognition, manifestation, evaluation and handling are presented for the exemplary interview
unknown “I do not know what will change”. Figure 19 provides the key. T refers to the interview day.
When I took the decision (X) about the method for the coaching (P), I did not know what will
have changed until P. At X I had applied scenario thinking (H) and had some ideas what
changes might happen but those were independent of each other. I only worked out the
concrete formulation “I do not know what will change” in the interview (T). At X it was still a
blindspot (B). In the interview (T) I recognised it as a clear unknown (C). At T I evaluated my
unknown as ambiguous (E), irrelevant (E) and potentially benefiting (E). I would deal
adaptively (H) with any changes that would have happened until P. I trust in my capacities
(H). At P it will clearly manifest what has changed.
Hence, the following code is applied:
• Time: X<T<P=M
• Time with classification of (un)known: XB, TC, MC
• Time with classification of (un)known, handling and evaluation: XBH, TCE, MCH
The reduced code of “I do not know what will change” would be: XBH<TCE<B=MCH
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
101
5.2.2
Interview Unknowns
A simplified version of the heuristic maps the following points in time (see Figure 19 for the key and
Figure 20 for the result): the decision (X), the manifestation of the unknown (M) and the reference
point of the decision (P). By definition, the non-knowledge mapped was at time X.
I did not know...
Timing
orders
Gradient
of
Unknown
at
Gradient
of
Known
at
Evaluation or
Handling
...
X, P, M
X
M
if study results are right or wrong
P<X
A
No
D prevented M
why the person was doing it
P<X
A
No/A
D made M irrelevant
specific formula to calculate it
P<X<M
A
D (or no)
nothing
P<X=M
no
D
actual access to land
X<P
C
No
D prevented M
how ecosystems will develop until
x
X<P<<M
A
A
the acceptance rate of locals
X<P<<M
A
D
is this good or bad?
X<P<<M
A
no/A/C/D
environmental impact
X<P<<M
B
C
consequences for the environment
X<P<<M
B
C
Sc: if there is life on planet X
X<P<<M
C
no/A/C/D
if strategies would have the
desired impact
X<P<<M
D
A
those who take home the message
X<P<M
A
A
it was unclear what intention x had
X<P<M
A
A/no
intentions of the government
X<P<M
A
A/no
how to activate them
X<P<M
A
A/C/D
scientific quality requirements
X<P<M
A
A/C/D
will the certification be successful
X<P<M
A
D
Sc: if results will improve
significantly
X<P<M A no/A/C
collaboration quality
X<P<M
C
A
if it would be positive or negative
X<P<M
C
D
who will be the partner
X<P<M
C
D
if we can change the legislation X<P<M
C
D
project design
X<P=M
A
A
attitude of the audience
X<P=M
A
A
if we will make too many
compromises
X<P=M
A
A
reaction of the council
X<P=M
A
A (or C)
if other people would also like it
and come
X<P=M
A
A/C/D
if it will work
X<P=M
A
C
If my proposal will be better
X<P=M
A
C
how clients will behave in terms of
X<P=M
A
D
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
102
x
If we have the financial resources
X<P=M
A
D
will stakeholders accept the final
set of x
X<P=M
B
D
what will change
X<P=M
B
C
who will be in the government
from x
X<P=M
C
D
security situation
X<P=M
N
A
internal power structure in the
village
X=P
A
No
complete range of criteria X=P
C
No
no M because not
useful, too complex
political and strategic thinking
X=P<M
A
A/no
how my team would react
X=P<M
A
C
Sc: for sure if x is the minister´s
name
X=P<M A D
details about other countries data
X=P<M
A
D/no
Sc: that my institution has
expertise on this
X=P<M B C
extent to which it contributes to
mitigation
X=P<M
C
D
target group
X=M<P
A
D
Sc: if the meeting with Y will take
place
Xp<M<X<P C delayed C
preliminary decision
Xp before M
Figure 20 shows the coded results for all 40 interview unknowns, five scenario unknowns (sc.) and the one interview
decision that did not have any unknown as it was taken by a directive (none). They are ordered according to timing.
Blocks of same colour have the same timing. Unknown and known provide the gradient of (non-)knowledge at X and M
respectively.
The following can be deduced from the results presented in Figure 20:
• Most decisions had a future reference point. Manifestation might either be a lot later, later
or happen at the reference point.
• Some decisions refer to the present. Manifestation can then either not happen or happen
later.
• In one decision the unknown manifested into a known before the reference point.
• Most unknowns are ambiguous. They can hence be approximated. They might commonly be
referred to as uncertainties or bounded unknowns.
• As the class of knowns is a future scenario, the exact class of manifestation can sometimes
not be assigned. A user should aim at giving the best guess.
• Possible (non-)handling (handling be H+, not-handling be H-) and evaluation (E)
o A decision can make M irrelevant (E-)
o A decision can prevent M (H+)
o Complexity can make M useless (H-)
o A preliminary decision can delay the real decision until manifestation (H+)
Those results presented above add to the following:
• Analysis of handling and evaluation would feed on the findings of non-knowledge literacy.
• Figure 21 gives instructions for the heuristic mapping of decision unknowns.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
103
Figure 21 gives a do-it-yourself instruction on how to heuristically map an unknown for a past or future decision.
5.2.3
Non-knowledge Literacy
Observed (non-)handling and evaluation in interviews indicate elements of non-knowledge literacy.
Handling was only observed in relation to time, as it was not specifically asked for handling of
unknowns but for how the decision was taken.
• The timing of a decision is a confirmed element of non-knowledge literacy.
• Evaluation can inform this handling.
Non-knowledge literacy cannot (yet?) prescriptively state how a non-knowledge literate person
would approach a certain situation. Principles have, however, been spelled out in this thesis that
could, through application and research, develop into guidelines. The confirmed element of non-
knowledge literacy is timing of the decision. The following findings of non-knowledge literacy
generated in the exploitation, conservation and release phase would have to be investigated and
possibly confirmed by coding or any other scientific heuristic.
Non-knowledge literacy is a cultural effort, a trial to strive between chaos and order, between the
unknown and the known. Non-knowledge literacy comprises changes in mind-set, behavioural
principles and rules and invites structural changes.
The coding is suggested as a tool to explore and create non-knowledge literacy. Application of the
code would make patterns and their associated handling and evaluation visible, as demonstrated in
above for decision timing. Those observed and evaluated handlings could be transferred into
principles of non-knowledge literacy. Analysis of those patterns would allow deriving preconditions
and competencies for evaluation and handling. Furthermore, every such application would allow the
user to better understand their implicit rules and heuristics. Such a better understanding would lift
them in consciousness so that they could be explicitly applied to justify decisions at work. First
principles for investigation would include handling of a non-knowledge literate person to
• Apply stopping rules
• Be familiar with and (un)consciously apply the non-knowledge map
• Consider the decision environment
• Consider the decision method (bias-conscious)
• Ignore existing knowledge at times
• Improve their non-knowledge-based reasoning abilities
• Know that ignorance based decision making works (recognition heuristic)
Use of the heuristic mapping tool for work unknowns
An individual that seeks to map a decision unknown would be called user. The individual could also
be a group as long as the combined ambiguity remains manageable. The user is asked to order four
points in time around the day of the mapping (T): the decision (X), the recognition of the unknown
(R), the manifestation of the according known (M) and the reference point of the decision (P). Then
at least X and M would be assigned a class of (non-)knowledge (as in Figure 19). Potential handling
and evaluation could then be assessed for the each point in time. Principles of non-knowledge
literacy might help.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
104
• Know that (non-)knowledge exists as a continuum
• Make more preliminary (tentative) decisions, but not forget they are preliminary
• Make their hypothesis (about the world) as late as possible
• Think past and future eventualities (counterfactual reasoning and scenarios)
5.3 Concluding Discussion
Reflection, definitions and application to interview unknowns were useful methods. Results propose
a heuristic non-knowledge map for decisions and confirm one element of non-knowledge literacy.
Further elements for non-knowledge literacy are suggested for further investigation. These results
will not be discussed in this thesis. With these results, this thesis has completed the four phases of
the panarchy.
The proposed third dimension of post-normal science invites reflections on how idiosyncrasies of
the post-normal scientists could be captured.
Several journals, which are accepted entities to set benchmarks for scientific work, nowadays allow
the use of the first person in scientific articles. No source is given for this sentence, and the
sentences referring to Heisenberg´s uncertainty principle (5.1.2) and Schroedinger´s cat (3.3.1.1).
This non-quoting should not be part of post-normal science. It has, however, been decided here that
this omission of quotation is functional as
(1) their acquisition might cost an unknown amount of time which is not available because of
the rules (heuristics) set by the university (hand in day after tomorrow),
(2) with the available time, results might be ambiguous for the author if no useful meta-study
can be found and capacities of the author might be overstrained with quantum mechanics,
(3) the uncertainty inherent in quantum mechanics can easily be perceived as ambiguous,
(4) it is hence heuristically relied on the knowledge that Heisenberg´s uncertainty principle and
Schroedinger´s cat are clear unknowns with societal recognition.
(5) There might not yet be a clear known (generated by meta-studies or more sophisticated
forms of societal consensus) in societal recognition about the use of the first person in
science.
(6) In code this could spell X-H1K+H2EJ: The decision (X) was not(-) handling (H) according to
known and clearly emerged rules (1) for knowledge acquisition (K) but to handle (H)
according to secondary rules designed for rule breaking or mistakes (2) which is making it
explicit (E) and justifying it (J). This code is a first proposal. It demonstrates how post-normal
scientific works could make their handling of limitations explicit. Stopping rules would have
to be developed so that such meta-justifications would not be lead ad absurdum. In scientific
works the need for justification of limitations is located in the methods and discussion
sections.
(7) One such stopping rule for post-normal science proposed by the above would be drilling
down to society´s clear unknowns. Clear unknowns should then not be left uncited but
undiscussed. If there are not enough clear unknowns to drill down to, an earlier stopping rule
had to be found.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
105
The use of first person singular in post-normal scientific work would be one step to allow
more idiosyncrasies and strengthen the new third dimension for post-normal science.
The third dimension of the post-normal science diagram should also be strengthened by
other elements.
Post-normal science can drill down to society´s clear unknowns, many of which have
been created and are being explored by science.
Science digs ever deeper into the fuzzy fringes of clearly emerged knowledge. Post-normal science
would bridge the gap from those fuzzy fringes to heuristic concepts that can be applied
transdisciplinary.
“Science is committed to rolling back the tide of ignorance in its relentless pursuit to provide
answers for all of the remaining unknown phenomena. It is no wonder that scientists have
eclipsed many other forms of enquiry and reasoning. The distinctive line of investigation
adopted in this field has encouraged the evolution of more sophisticated means of exploring
the environment in order to extend the knowledge boundaries of the known phenomena, and
to improve capabilities in dealing with problems and challenges. In his account, Karl Popper
(1966) describes this phenomenon as a trischema: The testing of a theory by experimentation
raises new problems or questions that prompt further investigation thus completing the
triangle and perpetuating the cycle of knowledge. This systematic approach has prompted
deeper and more detailed investigation into a rapidly growing list of disciplines. New frontiers
of knowledge have been opened up that have given rise to unimagined technology.
Notwithstanding, there are problems attached to ‘knowledge-wealth’ including logistical
issues of disseminating and sharing so much information with the wider society. Fear of the
unknown is the basis for man’s sense of insecurity and vulnerability. Ironically, in Popper’s
trischema, the pursuit of knowledge reveals the even larger gaps in human understanding of
the world, which in turn, feeds the paranoia of the unknown.” (Ibisch and Hobson, 2012a)
The panarchy, as an econical concept, might then conceptualise the cycle of non-knowledge on
which post-normal science be built. Like Popper´s trischema, it conceptualises the cycle of
(non-)knowledge. The table below attempts to conceptualise this.
Panarchy
Exploitation
Conservation
Release
Reorganisation
Popper´s
trischema:
the cycle of
knowledge
testing of a theory
by
experimentation
new problems or
questions
further
investigation
Understanding
created in this
thesis: the cycle
of non-
knowledge
Exploiting existing
concepts and
theories from
literature and
reflection
Conservation of
the emerged
concept or theory
through empirical
testing
Discussion
releases
conserved
concepts through
reflection
Reorganisation of
concepts into a
new concepts
that is applied to
conserved data
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
106
In accordance with the citation above, post-normal science would explicitly produce temporary and
tentative concepts to competently handle the unknown. It would then invite post-normal scientific
endeavours which attempt to create concepts which can be transdisciplinarily applied.
If Popper´s trischema describes science, post-normal science could be described by the
panarchy.
Post-normal science would produce temporary and tentative concepts in a cycle of non-
knowledge.
If knowledge and non-knowledge exist as a continuum, the cycle of non-knowledge
proposed for post-normal science should maybe called a non-knowledge aware cycle or
an oscillating between knowledge and non-knowledge.
Post-normal science would attempt to create concepts which can be applied
transdisciplinarily.
Questions for further exploration
• If the old non-knowledge map is a compass that points to directions within dimensions and
the new heuristic non-knowledge map codes unknowns, how could unknowns be mapped?
• Would a non-knowledge map not rather display patterns in coded unknowns and be
navigated by a compass? Would it not arrange inventoried unknowns?
• If the compass and the code can be applied to (non-)knowledge and (un)knowns, how could
those terms be displayed without brackets? Is removing the brackets enough or is a new
term and concept needed to gain significant societal recognition and finally manifestation?
6 Conclusion
The conclusion begins with short answers to the research questions. These are followed by
conclusions on the decision-centred exploration of non-knowledge. The insights are then
summarised into recommendations for decision makers. The thesis ends with an outlook and final
reflections.
6.1 Answers to Research Questions
6.1.1
What is non-knowledge literacy?
Compare 5.2.3: The timing of a decision is a confirmed element of how humans handle unknowns
and could hence become a principle of non-knowledge literacy. Several other principles of non-
knowledge literacy are suggested from literature and implicit handling (2.1.6 and 4.5). It is suggested
that the non-knowledge coding heuristic for decisions would facilitate its exploration and creation.
6.1.2
How can the non-knowledge map be applied to and improved for
decisions?
Figure 21: The non-knowledge map can be applied to decision unknowns. Those decision unknowns
must be clearly defined. The coding heuristic integrates all dimensions in a simple format.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
107
6.1.3
How volatile is the working environment of Global Change Managers?
3.3.3: Global Change Managers work in very different environments. The combined topical,
institutional and social volatility over 39 decisions tends to be medium.
6.1.4
Which heuristics do Global Change Managers use for work decisions?
4.2.2.1: No explicit heuristics could be found but many principles grounded in values and mind-set.
Those include but are not limited to adaptive management.
6.1.5
Which unknowns are common in work decisions?
3.3.2: Most unknowns are about circumstances, some about a decisions consequence and a lack of
complementary “how to approach the decision” knowledge. Most unknowns (5.2.2) were ambiguous
and might commonly be understood as bounded unknowns or uncertainties that can be
approximated.
6.1.6
Are Global Change Managers able to manage non-knowledge?
4.2.2.1: In 40 out of 41 decisions several unknowns could be easily named and had not significantly
hampered the decision. The handling of those unknowns was not explicit.
6.1.7
Are Unknowns Impeding Effective Decision Making?
4.2.2.2: From a sceptic but practical viewpoint how the decisions were taken seemed reasonable. If
unknowns do not impede effective decision making cannot be known. That one difficult case
surfaced in such a small sample with uncritically chosen decisions suggests it to remain an issue.
6.2 Conclusions on the Decision-centred Exploration of Non-knowledge
Non-knowledge plays a role in decision making. In fact, it was found in all but one decision. This one
decision was guided by a directive. The unknowns were mostly “unimpressive”. They appealed to
system 1 and were usually judged as unproblematic and common.
The econical approach allowed to exploit, conserve, release and reorganise several disciplinary and
transdisciplinary concepts. It proved useful in combination with interviews and reflections, which are
widely agreed methods across disciplines. Cookbook-style instructions and the panarchy were
applied a posteriori to structure the manuscript. This application of structure turned out to be
pattern finding. A fixed a priori structure might have prevented many results. This thesis has
provided an attempt to work with black boxes of imperfect and incomplete knowledge. The self-
referential cycle and the closely set boundaries of consideration allowed an inclusive exploration of
the topic.
A suitable, memorisable and easy to use non-knowledge mapping heuristic for decisions which
builds upon strong links between dimensions was derived from the multidimensional analysis of
unknowns.
The non-knowledge map is applicable to any unknown and any decision situation as long as both are
precisely defined. This reductionism is necessary to allow analysis. The decision can be situated in the
past, present or in the future. The do-it yourself instructions provided in Figure 21 guide potential
users in coding. Future decisions are more interesting in the sense that it remains an option to delay
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
108
the decision or work with preliminary decisions. Ambiguous (un)knowns are usually relevant for
global change related decision making.
Knowing the unknowable is impossible by definition.
One person´s unknown might be a known to others.
Today´s unknowns might be tomorrow´s knowns.
Yesterday´s knowns might today be unknown.
The unknowns analysed were “unimpressive”, this shows that no complicated decisions are needed
to find unknowns which are worth to be dealt with explicitly, but more complicated decisions might
reap additional results. However, this would be limited by individual decision making. At societal
scale, in democracies and times of flat hierarchies, many decisions are made not only in teams but by
various actors. This heuristic handling of decisions, and their related unknowns, by compromise,
negotiation or price allows chaos-to-order dynamics which would be difficult to grasp with the
proposed methods.
Unfortunately, no clear heuristics could be found in interviewee behaviour. This was partly limited by
the question “how did you decide” instead of “how did you deal with this non-knowledge”.
Additionally, it might need more sophisticated and probably disciplinary methods to systematically
exploit heuristics and their environment with the knowledge available.
The one-dimensional question-based exploration of dimensions of the compass non-knowledge map
is not recommended for further use. The questions confounded several interviewees and users and
many questioned their purpose. Thinking abstractly about unknowns is system 2 tiring. The heuristic
mapping against time uses commonplace concepts and terms that can be understood by system 1.
System 1 is quicker and often more coherent in its conclusions.
Throughout this thesis, an understanding of (un)knowns and (non-)knowledge has emerged that is
concluded here. While (non-)knowledge continues to embrace knowns, unknowns and blindspots,
(un)knowns are just one part of (non-)knowledge that can be actively handled as it is closely defined
and probably related to a specific situation.
Implicit handling of the unknown at work is common. Handling non-knowledge is an inherent
capacity of humans. It is suggested that the explicit recognition, analysis, and handling of non-
knowledge can provide benefits. Non-knowledge literacy could become a competence that is taught.
The initial statement about non-knowledge literacy is extended:
A competent understanding of non-knowledge is a key element of non-knowledge literacy. It can
be generated and enhanced with the heuristic non-knowledge map for decisions. Principles of
non-knowledge literacy include heuristics, mind-set as well as environmental variables. It
considers frameworks of adaptive, integrative and flexible management and robust or resilient,
participatory, proactive, precautionary and preventive strategies. Together with a bias-conscious
use of heuristics and stopping rules it can facilitate non-knowledge based decision making.
This rough understanding of non-knowledge literacy already points to key principles, but leaves the
concept open for further exploration. That non-knowledge literate individuals are happier and that
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
109
the decisions they take converge to a more effective implementation at institutional level remains a
postulate. This postulate will be further explored.
6.3 Recommendations for Decision Makers
The heuristic non-knowledge map could be applied quite reliably in practical contexts and would gain
from applied and disciplinary perspectives. It is suggested that an explicit handling of unknowns be
learnt. Cultural stability has taught humans for too long that the world is knowable. Stepping back,
widening boundaries of consideration and allowing more non-knowledge into professional contexts
would be a necessary and humble first step.
“We just don’t want to just survive uncertainty, to just about make it. We want to survive
uncertainty and, in addition—like a certain class of aggressive Roman Stoics—have the last
word. The mission is how to domesticate, even dominate, even conquer, the unseen, the
opaque, and the inexplicable.” (Taleb, 2012)
Decisions at work are concrete moments for the recognition and manifestation of (un)knowns. This
should be made more explicit by allowing their active evaluation and competent handling. It is
suggested that decision makers at any level attach greater value to the unknown and consider
justifying their decisions with all their (un)knowns.
By questioning and challenging one´s own handling of unknowns and by inspiration and theoretical
background, principles of non-knowledge literacy could be created. Non-knowledge literacy is a step
towards competent non-knowledge based decision making at work. A non-knowledge based
decision making at work would not require radical change. Scenario thinking, assumptions about the
future and adaptive management, together with multiple other tools to handling the unknown are
commonly used. It is recommended to keep a humble and open mind and to allow discussing
complexity and chaos. Action, however, would have to be guided by clear and adaptive heuristics,
which finally apply reductionism and generate order to provide a secure environment for humans to
work and decide. Human well-being in times of global change, knowledge overload and constant
availability, remains the key dimension.
This thesis has touched clear scientific unknowns such as dark matter and quantum mechanics and
then used intelligible definitions, specifications and heuristics to lift these unknowns into practical
transdisciplinary application. This proceeding is suggested for post-normal scientific works.
6.4 Outlook and Final Reflections
This thesis might provide one example where science has created an overarching heuristic for
management in times of rapid global change.
“Science and human intuition are useful templates on which to build a new post-normal
science infused with risk-aversion management and principles of non-knowledge. The
essentials of a new school of science based on principles of non-knowledge include a
classification system and a set of descriptors. Once established, a fresh perspective on
principles and practices of sustainability can then be formulated.” (Ibisch and Hobson,
2012a).
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
110
This thesis extended post-normal science into a third dimension. If this was to be implemented
across the field of post-normal science, it would allow integrating intuition or its explicit derivatives.
It would then also comprise all three clusters found in (un)knowns: consequences, circumstances and
complementarity. Suggestions are to be made how the points outside the red realm on the post-
normal science diagram can be treated. This is probably not non-science.
The panarchy has allowed structuring this post-normal scientific work. It is suggested that the
panarchy could describe the cycle of non-knowledge which post-normal science would use to break
new ground. This new ground would consist of explicitly temporary and tentative concepts for
transdisciplinary application.
It might turn out that, at societal scale, similar or more precise models exist for the systematic
exploration and classification of decision unknowns and the implementation and structuring of post-
normal science. If they do not exist now, they might come into existence in the future. In this case, it
is hoped that the understanding and taxonomy generated in this thesis can advance the societal
dialogue. This thesis would recommend that a common definition of non-knowledge and a clear
taxonomy should not yet be provided. Retaining uncertainties and divergent classifications for some
time might yield significant benefits, as adaptive management and preliminary decision making
demonstrate. One taxonomy which is useful for a heuristic and decision related classification of
(un)knowns, their recognition and manifestation has been developed. Handling and evaluation
extend into non-knowledge literacy and will be further demarcated.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
111
VI. Bibliography
Barker, T., Bashmakov, I., Bernstein, L., Bogner, J. E., Bosch, P. R., Dave, R., Davidson, O. R., Fisher, B.
S., Gupta, S., Halsnæs, K., Heij, G. J., Kahn Ribeiro, S., Kobayashi, S., Levine, M. D., Martino, D. L.,
Masera, O., Metz, B., Meyer, L. A., Nabuurs, G.-J., Najam, A., Nakicenovic, N., Rogner, H.-H., Roy, J.,
Sathaye, J., Schock, R., Shukla, P., Sims, R. E. H., Smith, P., Tirpak, D. A., Urge-Vorsatz, D. and Zhou, D.
(2007) Technical Summary, In Climate Change 2007: Mitigation. Contribution of Working Group III to
the Fourth Assessment Report of the Intergovernmental Panel on Climate Change [B. Metz, O. R.
Davidson, P. R. Bosch, R. Dave, L. A. Meyer (eds)], Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA, [online] Available at: http://www.ipcc.ch/pdf/assessment-
report/ar4/wg2/ar4-wg2-chapter3.pdf (Accessed 18 October 2015).
Baron, J. (2000) Thinking and Deciding, 3d ed, New York, Cambridge University Press.
Baron, J. (2008) Thinking and Deciding, 4th ed, New York, Cambridge University Press.
Berry, W. (ed.) (2008) The Way of Ignorance, In The Virtues of Ignorance Complexity, Sustainability
and the Limits of Knowledge, The University Press of Kentucky.
Breuer, F. (2010) Wissenschaftstheoretische Grundlagen qualitativer Methodik in der Psychologie, In
Mey, G. and Mruck, K. (eds.), Handbuch Qualitative Forschung in der Psychologie, VS Verlag für
Sozialwissenschaften, pp. 35–49.
von Clausewitz, C. (1832) Vom Kriege, Berlin, Dümmlers Verlag, [online] Available at:
http://www.wissensnavigator.com/documents/CarlvonClausewitzVomKriege.pdf (Accessed 30 July
2015).
Dean, R. H. (ed.) (2008) Optimizing Uncertainty, In The Virtues of Ignorance Complexity, Sustainability
and the Limits of Knowledge, The University Press of Kentucky.
Dessai, S. and Wilby, R. (2011) How can developing country decision makers incorporate uncertainty
about climate risks into existing planning and policymaking processes, World Resources Report
Uncertainty Series. Washington, DC: World Resources Institute. http://www. worldresourcesreport.
org/decision-making-indepth/managing-uncertainty.(15 September 2012), [online] Available at:
http://www.wri.org/sites/default/files/uploads/wrr_dessai_and_wilby_uncertainty.pdf (Accessed 18
October 2015).
Eisenhardt, K. (2015a) Kathleen Eisenhardt on Simple Rules, Max Planck Institute for Human
Development, [online] Available at: https://www.youtube.com/watch?v=pO2QpGLibjQ (Accessed 12
August 2015).
Eisenhardt, K. (2015b) Simple Rules Q&A with Kathy, [online] Available at:
http://www.simplerulesbook.com/qa-with-kathy.html (Accessed 13 October 2015).
Eisenhardt, K. M. (1989) Building theories from case study research, Academy of management
review, 14(4), pp. 532–550.
Eisenhardt, K. M. and Graebner, M. E. (2007) Theory building from cases: opportunities and
challenges, Academy of management journal, 50(1), pp. 25–32.
Faculty of Forestry and Environment of the University of Applied Sciences, Eberswalde (2009)
Curriculum and Module Description Master study programme Global Change Management effective
from WS 2009/2010, [online] Available at: http://hnee.de/en/Programmes/Master-degree/Global-
Change-Management/students/Students-K1765.htm.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
112
Faculty of Forestry of the University of Applied Sciences, Eberswalde (2006) Curriculum and Module
Description International Master Study Programme Global Change Management (M.Sc.)(effective
from WS 2006/07), [online] Available at: http://hnee.de/en/Programmes/Master-degree/Global-
Change-Management/students/Students-K1765.htm.
Freudenberger, L., Schluck, M., Vega, E. A., Sommer, H., Cramer, W., Barthlott, W. and Ibisch, P. L.
(2010) A View on Global Patterns and Interlinkages of Biodiversity and Human Development, In
Interdependence of Biodiversity and Development Under Global Change CBD Technical Series No. 54,
[online] Available at:
http://sa.indiaenvironmentportal.org.in/files/interdependence%20of%20biodiversity.pdf (Accessed
27 October 2013).
Funtowicz, S. and Ravetz, J. (2003) Post-normal science, International Society for Ecological
Economics (ed.), Online Encyclopedia of Ecological Economics at http://www. ecoeco.
org/publica/encyc. htm.
Geiger, L., Kreft, S. and Ibisch, P. L. (2012) Reducing Blindspots: MARISCO, a Planning Approach that
Integrates Risk Management into Biodiversity Conservation, In Global Change Management:
Knowledge Gaps, Blindspots and Unknowables, Baden-Baden, Nomos.
Gigerenzer, G. (2007a) Bauchentscheidungen: die Intelligenz des Unbewussten und die Macht der
Intuition, Bertelsmann.
Gigerenzer, G. (2007b) Gut Feelings: The Intelligence of the Unconscious, Viking.
Gigerenzer, G. (2006) Heuristics, In Heuristics and the Law, MIT Press.
Gigerenzer, G. (1991) How to make cognitive illusions disappear: Beyond “heuristics and biases”,
European review of social psychology, 2(1), pp. 83–115.
Gigerenzer, G. and Brighton, H. (2009) Homo Heuristicus: Why Biased Minds Make Better Inferences,
Topics in Cognitive Science, 1(1), pp. 107–143.
Gigerenzer, G., and C. Engel. Heuristics and the Law. MIT Press, 2006.
Gigerenzer, G. and Todd, P. M. (1999) The Research Agenda, In Simple Heuristics that Make us Smart,
New York, Oxford University Press.
Gigerenzer, G., Todd, P. M. and The ABC Research Group (1999) Simple Heuristics that Make us
Smart, New York, Oxford University Press.
Giles, L. (1910) The art of war by Sun Tzu, El Paso Norte, [online] Available at:
http://jstrack.org/suntzu/Sun_Tzu-The_Art_of_War_Lionel_Giles.pdf (Accessed 30 July 2015).
Gilovich, T., Griffin, D. and Kahneman, D. (eds.) (2002) Heuristics and Biases The Psychology of
Intuitive Judgment, Cambridge University Press.
Gunderson, L. H. and Holling, C. S. (2002) Panarchy Understanding Transformations in Human and
Natural Systems, Washington Covelo London, Island Press.
Günther-Dieng, Guericke, M., Spathelf, P., Schultz, A., Welp, M., Nowicki, C. and Jahn, T. (2014)
Anhang zum Bericht des Fachbereichs für Wald und Umwelt der Hochschule für nachhaltige
Entwicklung Eberswalde (FH) zur Re-Akkreditierung der Bachelorstudiengänge Forstwirtschaft (B.Sc.)
und International Forest Ecosystem Management (B.Sc.) sowie der Internationalen
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
113
Masterstudiengänge Forest Information Technology (M.Sc.) und Global Change Management (M.Sc.),
unpublished.
Günther-Dieng, Guericke, M., Spathelf, P., Schultz, A., Welp, M., Nowicki, C. and Jahn, T. (2015)
Stellungnahme des Fachbereichs für Wald und Umwelt der Hochschule für nachhaltige Entwicklung
Eberswalde zum ASIIN-Gutachterbericht zur Re-Akkreditierung der Studiengänge Forstwirtschaft
(B.Sc.); International Forest Ecosystem Management (B.Sc.), Forest Information Technology (M.Sc.)
und Global Change Management (M.Sc.), unpublished.
Günther-Dieng, K., Guericke, M., Spathelf, P., Schultz, A., Welp, M., Nowicki, C. and Jahn, T. (2014)
Bericht des Fachbereichs für Wald und Umwelt der Hochschule für nachhaltige Entwicklung
Eberswalde (FH) zur Re-Akkreditierung der Bachelorstudiengänge Forstwirtschaft (B.Sc.) und
International Forest Ecosystem Management (B.Sc.) sowie der Internationalen Masterstudiengänge
Forest Information Technology (M.Sc.) und Global Change Management (M.Sc.)., unpublished.
Halsnæs, K., Shukla, P., Ahuja, D., Akumu, G., Beale, R., Edmonds, J., Gollier, C., Grübler, A., Ha
Duong, M., Markandya, A., McFarland, M., Nikitina, E., Sugiyama, T., Villavicencio, A. and Zou, J.
(2007) Framing issues, In Climate Change 2007: Mitigation. Contribution of Working Group III to the
Fourth Assessment Report of the Intergovernmental Panel on Climate Change [B. Metz, O. R.
Davidson, P. R. Bosch, R. Dave, L. A. Meyer (eds)], Cambridge University Press, Cambridge, United
Kingdom and New York, NY, USA, [online] Available at: http://www.ipcc.ch/pdf/assessment-
report/ar4/wg2/ar4-wg2-chapter3.pdf (Accessed 18 October 2015).
Hammer, Ø., Harper, D. A. T. and Ryan, P. D. (2001) PAST: Paleontological Statistics Software Package
for Education and Data Analysis., Palaeontologia Electronica 4(1): 9pp., 4((1)), p. 9pp.
Hobson, P. and Ibisch, P. L. (2010) Strategic Sustainable Development: A Synthesis towards
Thermodynamically Efficient Systems and Post-Normal Complex Systems Management, In
Interdependence of Biodiversity and Development Under Global Change CBD Technical Series No. 54,
[online] Available at:
http://sa.indiaenvironmentportal.org.in/files/interdependence%20of%20biodiversity.pdf (Accessed
27 October 2013).
Hobson, P. R. (2015) RE: Trial run results - eight non-knowledge maps and appointment request,
unpublished email.
Hobson, P. R. and Ibisch, P. L. (2013) “Forest econics:” mimicking processes and patterns in old
growth forest to promote sustainable forestry under global change, In In: Ministry for Ecology and
Natural Resources of Ukraine & Carpathian Biosphere Reserve (eds.) 2013. Primeval and ancient
beech forests of Europe: problems of protection and sustainable use. Proceedings of the International
Conference, Rakhiv, September 16-22, 2013. 66-76.
Hobson, P. R. and Ibisch, P. L. (2012) Learning from Nature for Sustainability: An Econical Approach to
(Non-)Knowledge Management, In Global Change Management: Knowledge Gaps, Blindspots and
Unknowables, Baden-Baden, Nomos.
Hoffrage, U. and Hertwig, R. (1999) Hindsight Bias: A Price Worth Paying for Fast and Frugal Memory,
In Simple Heuristics that Make us Smart, New York, Oxford University Press.
Ibisch, P. L. (2010) Global change management: eine systemische Utopie der Nachhaltigkeit im
Angesicht der Apokalypse, In ‘Simplizistische Lösungen verbieten sich’. Zur internationalen
Zusammenarbeit im 21. Jahrhundert. Festschrift zu Ehren von Professor Uwe Holtz, Nomos, Baden-
Baden, E. Deutscher & H.Ihne (eds.).
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
114
Ibisch, P. L. (2015) Meeting, Discussion of Master Thesis, personal communication, unpublished.
Ibisch, P. L., Geiger, L. and Cybulla, F. (eds.) (2012) Global Change Management: Knowledge Gaps,
Blindspots and Unknowables, Baden-Baden, Nomos.
Ibisch, P. L. and Hobson, P. R. (2012a) Blindspots and Sustainability under Global Change: Non-
knowledge Illiteracy as a Key Challenge to a Knowledge Society, In Global Change Management:
Knowledge Gaps, Blindspots and Unknowables, Baden-Baden, Nomos.
Ibisch, P. L. and Hobson, P. R. (eds.) (2014) MARISCO: Adaptive MAnagement of vulnerability and RISk
at COnservation sites: A guidebook for risk-robust, adaptive and ecosystem-based conservation of
biodiversity, Eberswalde, Centre for Econics and Ecosystem Management.
Ibisch, P. L. and Hobson, P. R. (2012b) Working beyond evidence and knowledge boundaries to
achieve sustainable development in a rapidly changing world, prepared for the 2012 Berlin
Conference on Evidence for Sustainable Development, unpublished.
Ibisch, P. L., Vega, E. A. and Herrmann, T. M. (eds.) (2010) Interdependence of Biodiversity and
Development Under Global Change, CBD Technical Series No. 54, [online] Available at:
http://sa.indiaenvironmentportal.org.in/files/interdependence%20of%20biodiversity.pdf (Accessed
27 October 2013).
Jackson, W. (ed.) (2008) Toward an Ignorance-based worldview, In The Virtues of Ignorance
Complexity, Sustainability and the Limits of Knowledge, The University Press of Kentucky.
Kahneman, D. (2011) Thinking, Fast and Slow, New York, Farrar, Straus and Giroux.
Kahneman, D. and Klein, G. (2009) Conditions for Intuitive Expertise a Failure to Disagree, American
Psychologist, 10/2009, pp. 515–526.
Klein, G. A. (1999) Sources of Power: How People Make Decisions, 2nd ed, Cambridge, Massachusetts,
London, England, MIT Press.
Kundzewicz, Z. W., Mata, L. J., Arnell, N. W., Döll, P., Kabat, P., Jiménez, B., Miller, K. A., Oki, T., Sen,
Z. and Shiklomanov, I. A. (2007) Freshwater resources and their management, In Climate Change
2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth
Assessment Report of the Intergovernmental Panel on Climate Change, M.L. Parry, O.F. Canziani, J.P.
Palutikof, P.J. van der Linden and C.E. Hanson, Eds., Cambridge University Press, Cambridge, UK, 173-
210, [online] Available at: http://www.ipcc.ch/pdf/assessment-report/ar4/wg2/ar4-wg2-
chapter3.pdf (Accessed 18 October 2015).
Lawton, J. H. (2007) Ecology, politics and policy, Journal of Applied Ecology, 44(3), pp. 465–474.
Librevis (2006) Zusammenfassung von Das Einmaleins der Skepsis, [online] Available at:
http://www.univie.ac.at/soziologie-statistik/pflege/UE/GG-Das_Einmaleins_der_Skepsis.pdf
(Accessed 8 July 2015).
Martignon, L. and Hoffrage, U. (1999) Why Does One-Reason Decision Making Work? A Case Study in
Ecological Rationality, In Simple Heuristics that Make us Smart, New York, Oxford University Press.
McKinsey&Company (2015) Women in the Workplace 2015, [online] Available at:
http://womenintheworkplace.com/ui/pdfs/Women_in_the_Workplace_2015.pdf?v=5 (Accessed 23
October 2015).
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
115
Nugent, C. (2008) Ignorance and Know-how, In The Virtues of Ignorance Complexity, Sustainability
and the Limits of Knowledge, The University Press of Kentucky.
Perry, R. (2008) Ignorance - an Inner Perspective, In The Virtues of Ignorance Complexity,
Sustainability and the Limits of Knowledge, The University Press of Kentucky.
Peterson, A. L. (2008) Ignorance and Ethics, In The Virtues of Ignorance Complexity, Sustainability and
the Limits of Knowledge, The University Press of Kentucky.
Rieskamp, J. and Hoffrage, U. (1999) When Do People Use Simple Heuristics, and How Can We Tell?,
In Simple Heuristics that Make us Smart, New York, Oxford University Press.
Rumsfeld, D. H. (2002) Transcript: DoD News Briefing - Secretary Rumsfeld and Gen. Myers, [online]
Available at: http://www.defense.gov/transcripts/transcript.aspx?transcriptid=2636 (Accessed 12
August 2015).
Slovic, P., Finucane, M., Peters, E. and MacGregor, D. (2008) The Affect Heuristic, In Heuristics and
biases: the psychology of intuitive judgment, 7. printing, Cambridge, Cambridge Univ. Press.
Steffen, W., Grinevald, J., Crutzen, P. and McNeill, J. (2011) The Anthropocene: conceptual and
historical perspectives, Philosophical Transactions of the Royal Society A: Mathematical, Physical and
Engineering Sciences, 369(1938), pp. 842–867.
Sutherland, W. J., Armstrong-Brown, S., Armsworth, P. R., Tom, B., Brickland, J., Campbell, C. D.,
Chamberlain, D. E., Cooke, A. I., Dulvy, N. K., Dusic, N. R. and others (2006) The identification of 100
ecological questions of high policy relevance in the UK, Journal of applied ecology, 43(4), pp. 617–
627.
Talbott, S. (2008) Toward an Ecological Conversation, In The Virtues of Ignorance Complexity,
Sustainability and the Limits of Knowledge, The University Press of Kentucky.
Taleb, N. N. (2012) Antifragile: Things That Gain from Disorder, New York, Random House.
Taleb, N. N. (2004) Fooled by Randomness The Hidden Role of Chance in Life and in the Markets,
Second Edition, New York, Texere.
Taleb, N. N. (2008) The Black Swan: The Impact of the Highly Improbable, New York, Random House
Trade Paperbacks.
Todd, P. M. and Gigerenzer, G. (1999) What We Have Learned (So Far), In Simple Heuristics that
Make us Smart, New York, Oxford University Press.
Tversky, A. and Kahneman, D. (1974) Judgement under Uncertainty: Heuristics and Biases, Science,
185(4157), pp. 1124–1131.
Tversky, A. and Koehler, D. J. (2002) Support Theory: A Nonextensional Representation of Subjective
Probability, In Heuristics and Biases The Psychology of Intuitive Judgment, Cambridge University
Press.
UNFCCC (1992) United Nations Framework Convention on Climate Change, [online] Available at:
http://www.official-documents.gov.uk/document/cm28/2833/2833.pdf (Accessed 17 February
2015).
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
116
Vitek, B. and Jackson, W. (2008a) Taking Ignorance Seriously, In The Virtues of Ignorance Complexity,
Sustainability and the Limits of Knowledge, The University Press of Kentucky.
Vitek, B. and Jackson, W. (eds.) (2008b) The Virtues of Ignorance Complexity, Sustainability and the
Limits of Knowledge, The University Press of Kentucky.
Welling, M. (2008) Das Johann Heinrich von Thünen-Institut ist Partner des Studiengangs „Global
Change Management“ Erweiterung des Kooperationsvertrags mit der Fachhochschule Eberswalde,
Johann Heinrich von Thünen-Institut (vTI) Bundesforschungsinstitut für Ländliche Räume, Wald und
Fischerei, [online] Available at: http://hnee.de/en/Programmes/Master-degree/Global-Change-
Management/about-GCM/news-events/press-releases/Press-releases-K2411.htm.
Witte, M. H., Crown, P., Bernas, M. and Witte, C. L. (2008) Lessons learned from ignorance, In The
Virtues of Ignorance Complexity, Sustainability and the Limits of Knowledge, The University Press of
Kentucky.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
117
VII. Annex A
Content of Annex A
1 Board Game Map
2 About Global Change Management (MSc)
3 Trial Run
4 Exploring (the) Non-knowledge (map) II
1. Board Game Map
Figure 22 shows the non-knowledge map transferred into a board game. Additionally the user needs cards with both
ends of the dimensions (front and back) which have to be placed on the schedule. A draft version of the game
instructions is written but not published.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
118
2. About Global Change Management (MSc)
A Global Change Manager is a person who has successfully completed the International Master Study
Programme “Global Change Management (MSc)” at Eberswalde University for Sustainable
Development (before 2010 the University of Applied Sciences, Eberswalde). The GCM study
programme started in September 2006 out of the need to translate knowledge generated about
global environmental changes into applied concepts. An applied link was established from the start
by the inclusion of a board and five active partner organizations.
A German article titled “Global change management: a systemic utopia in view of the apocalypse”
(Ibisch, 2010) illustrates the need for global change management. Accordingly, Global Change
Management means not allowing global change to just happen but to try and transform an
unmanageable amount of knowledge and non-knowledge into a different direction. This is according
to the utopia that mankind as a common entity can reach a sufficient holarchical intelligence to
transform the complex anthroposystem and wide parts of the Earth system in the sense of
sustainability. This is also based on the observation that human culture has developed a shell in
which they are largely self-referential and feeling safe, as if they did not depend on the ecosystem
which dramatically changes in its functionality. It asks if mankind is able to purposefully influence,
stop or reverse the developing of systems and, can we be more conscious and strategic to learn a
truly sustainable handling of the Earth system. It asks if Global Change Management can work or is
just a chronical overestimation of human possibilities.
Global Change Management is a transdiscipline for a change management across all hierarchical
planes of social systems. It includes adaptation to unavoidable global change processes and effects
without neglecting the aversion of dangerous changes. Mitigation and deceleration of these
processes is of primary importance. Global Change Management means conceptual and operative
changes across scales of complex social systems. It accepts the necessity of acting and thinking
simultaneously at all spatial scales. GCM is not about saving the world but about avoiding the short-
term apocalypse of human civilisation. (Ibisch, 2010)
“From the beginning it became clear that any efforts of managing causes and consequences of global
change would have to deal with uncertainty” (Ibisch, Geiger and Cybulla, 2012). A heterogeneous
group of students is taught transdisciplinary; systems thinking, adaptive management and proactive
risk management as well as strategic planning provide conceptual frameworks (Günther-Dieng et al.,
2015).
An analysis carried out for the ASIIN accreditation of the programme (Günther-Dieng et al., 2014)
describes occupational areas of alumni and derives the necessity of the programme. Accordingly,
science and politics need people who can plan and implement concrete mitigation and adaptation
measures. Whereas comparable study programmes are usually focused on basic natural scientific
research. These generate an incredible amount of information but practical application could not yet
take advantage of. The aim of GCM is to act and learn adaptively. Alumni could work in research,
global environmental politics, national environmental politics, insurances, organisational
development or political consulting. The results of a survey which was answered by 31 GCM alumni
shows that 29% work in scientific institutions and universities, 23% in the economy and in
administration (22%) as well as 13% each in NGOs or as freelancers. Fields are solution-oriented,
practical and political.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
119
The unique programme is a consecutive offer for forestry, land use and conservation as well as other
natural scientific, ecological and environmental studies. However, also career changers are admitted
(Günther-Dieng et al., 2014). Students have complete a large variety of bachelor programmes, e.g.
biology, forestry, education, philosophy, business administration, economics, international relations
etc. Many have worked before they started studying Global Change Management. Hence also the
age of students at start ranges from early twenties to mid-forties.
GCM has an annual capacity of 25 or 26 students (compare Figure 24). The number of applicants in
2006 and 2007 was 22 and 27 with 14 and 13 first-year students respectively. From 2008 there are
annually 41 to 99 applicants and 20 to 26 first-year students (compare Figure 23). In 2008 eight
students finished the study programme and two in 2009. From 2010 to 2014 the annual number of
new alumni ranges from 12 to 19. From 2006 to 2012 the average duration to finish GCM was 5.5
semesters. In November 2014, there were 86 alumni. Admission is guaranteed to at least 40%
international (means non-German) students. There were no international first year full-time students
in 2006, one in 2007 and 2009, two in 2008. From 2009 to 2013, the number of international first
year students steadily grew to 13. About one third of total students have spent their research
semester outside of Germany. (Günther-Dieng et al., 2014)
Figure 23 First-year GCM students from 2006 to 2014, figure adapted from Günther-Dieng et al. (2014)
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
120
Figure 24 GCM students from 2006-2014, figure adapted from Günther-Dieng et al. (2014)
Along with the university, the study programme’s partner organizations are the development and
environment NGO Germanwatch e.V., the German development agency giz (gtz until September
2010), the reinsurance company Munich Re, the nature conservation NGO NABU, and the Potsdam
Institute for Climate Impact Research PIK. Since 2008, the Federal Research Institute for Rural Areas,
Forestry and Fisheries Johann Heinrich von Thünen Institute, vTI, is also partner of the study
programme (Welling, 2008). These institutions actively shape the content of the study programme by
teaching certain modules or their parts, hosting research projects and theses as well as occasionally
becoming the employer of alumni. Guest lecturers (often the study programme’s or university’s
alumni) frequently teach single classes or blocks within modules.
There was an initial version of the curriculum of the study programme effective from winter
semester 2006/2007, a new version effective from winter semester 2009/2010 and a current change
took place which will be effective from winter semester 2015/2016. Alumni interviewed in this thesis
have studied under the two old curricula.
In the first curriculum, first semester mandatory modules were “Physical fundamentals of global
change processes” (taught at PIK), “Global change ecology”, “Socioeconomic and institutional
dimensions of global change” and “Management of conservation organizations and lobbying I”
(taught at NABU). Second semester mandatory modules included “Moderation, negotiation &
conflict management”, “Communication of global change issues”, “Global change, risk management
& insurance industry” (taught by Munich Re), “Conservation management & land-use management
under global change”, “Applied climate policy” (taught by Germanwatch) and “Forest management &
adaptation to global change”. In the third semester, students were to carry out a “practical research
project on a specific topic related to the study programme’s content”. This was accompanied by a
research colloquium with online discussions. In the fourth semester, students were requested to
write their Master thesis and participate in the module “Designing climate change mitigation
projects”. Elective modules included ecology, remote sensing, statistics and data, modelling,
lobbying, forestry topics, project management, globalization and international cooperation (taught
by GTZ). (Faculty of Forestry of the University of Applied Sciences, Eberswalde, 2006)
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
121
In the second curriculum, effective from Winter Semester 2009/2010, first semester mandatory
modules were “Conditions and tools: change management” (partly taught by NABU), “Objects and
dynamics: global systems analysis” (partly taught by PIK), and in the second semester “Response
strategies: adaptation to global change” (partly taught by Munich Re) and “Response strategies:
mitigation of global change” (partly taught by Germanwatch). A “scientific internet colloquium”
accompanied the third semester’s “Research project” and a “Research colloquium” the “Master
thesis and defence” in the fourth semester. Elective modules covered similar topics as in the first
curriculum. (Faculty of Forestry and Environment of the University of Applied Sciences, Eberswalde,
2009)
The third curriculum, effective from winter semester 2015/16 is structured along similar lines,
expressed as topical and methodological expertise in systems theory, natural and social sciences,
decision making and responsibility taught in terms of change management, skills in presentation and
communication, ability to work in teams, moderation and conflict resolution as well as intercultural
social and interpersonal skills through international and interdisciplinary projects and research work.
(Günther-Dieng et al., 2014)
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
122
3. Trial Run
The trial run questionnaire (annex VIII.2) was built around the non-knowledge map. It was based on
the literature consulted so far. Four interviews were carried out personally with GCM students
(GCM2013) on decision situations from their internships at 7 and 8 May 2015, they took roughly 30
minutes each. The questionnaire asked for basic information, contained the post-normal science
diagram (Funtowicz and Ravetz, 2003) and two non-knowledge maps for unknowns encountered in
two decision situations. For development and background of the questionnaire see annex VIII.10.
Results were displayed by bar charts and webs respectively for the non-knowledge map and by dots
(or different forms) on the post-normal science diagram. Personal data were not evaluated.
Results
No significant pattern could be recognized in the eight decisions gained from the trial interviews,
compare Figure 25 and Figure 26. Main result of the trial interviews was the fact that the general
interview design and questionnaire work out, that analysis of results needs better methods and that
some questions and options have to be improved.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
123
Figure 25: All mapped non-knowledge from trial run interview decisions was recognized, related to the future or present
and had a threat potential. For some the geographical or social distribution dimensions did not apply. There are more
bars in the black “more sustainability relevant” part of the map. Figure 10 provides the key for the gradients.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
124
Figure 26: The decisions cluster strongly in the second and partly in the third realm of the diagram. In three of four cases
the first and second decision are placed in a similar place on the diagram and apply the same decision "method".
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
125
4. Exploring (the) Non-knowledge (map) II
After the interviews, aspects of the analysed dimensions and the classification of unknowns
remained unclear. Hence, additional knowns and unknowns were mapped. Parts of this tinkering are
displayed in this section. Results of these trials and reflections are presented in the main text. The
dimensions are in ordered according to their importance as perceived to the user of the map. This is
a working document; it real-time tracked the process how the result was generated; this is reflected
in form and language.
Map “I do not know how to construct a new non-knowledge map”
Reflections
on
Dimensions
- I recognised I do not know how to construct a new non-knowledge map
(recognition)
- This extends through the present (temporality) but the non-knowledge can
end soon: as it is a how to decision, knowledge and decision are the same
thing. If one option is created and chosen (if the if to decision is to take the
first generated map). If more options are created (knowable) it becomes a
what exactly decision.
- I am quite sure there is no non-knowledge map out there (in the media) as I
imagine it to be (geography / locality)
- I am rather sure there is nobody else who knows (social distribution) -> and
if there is, it is now not available as I do not know how to find it or did not try
hard enough to find it because of imaginary, time or resource limits – and if
my supervisor knows (more) he wants me to work it out or my current
intellectual capacities are too limited to understand
and apply his
explications.
- It is unintentional (-> locality, social distribution) but I have the intention to
make it knowable -> how can I make it knowable? (literacy)
- It is relevant (relevance) because I want to have it (goal). I could change or
abandon my goals and it would not be relevant (literacy?).
- It is potentially benefiting (threat potential)
- It does not feel ambiguous for me
Result and
Discussion
Hence I know about my non-knowledge: I have the chance to turn it into knowledge
(knowable by definition). I know I could investigate more to get more inspiration (by
information that is socially and geographically distributed) but I decided not to do it
actively (stopping rule/a heuristic). As it is relevant and potentially benefiting for me I
decided to try to turn my non-
knowledge into knowledge (by thinking, tinkering,
trying, asking, discussing…). This process is still ongoing.
Conclusion
Using the non-knowledge map requires analysing (objectively) and deciding
(subjectively attaching value) for ways how to deal with the non-knowledge.
The following tasks aimed at testing the categories of mapped non-knowledge (value, future, others,
information.). They were to illustrate the main impediment that the non-knowledge would become
knowledge and were functional as they were mutually exclusive but doubt rose they would be
functional in general. After this task they were dismissed.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
126
Map a “future” related aspect of non-knowledge from the interviews from the perspective of the
interviewee “if strategies would have the desired impact” – a what exactly decision.
Reflections
on
Dimensions
- Recognised.
- Future temporality.
-
I decided not to acquire more information that might help me to
approximate better (exploit the social distribution and geography of
knowledge). Apart from that it is unintentional.
- I perceive a threat potential.
- It is ambiguous
because I have either not defined or cannot measure the
exact desired impact (it is too complex and I do not want to be too
prescriptive and reductionist -> intentionality
). I hence understand it as
unknowable. If I knew it would be relevant.
Conclusion
The impediment of knowledge is here temporality (the ambiguity could be solved), as
the non-knowledge is about the consequences of my decision.
Map an “information” related aspect of non-knowledge from the interviews from the perspective
of the interviewee “scientific quality requirements” – a what exactly decision.
Reflections
on
Dimensions
- Recognised.
- Extends through the present.
- For me those are highly ambiguous
so I decided not to acquire more
information that might help me to approximate better (exploit the social
distribution and geography
of knowledge). Hence I perceive it as
unintentional. It was now unknowable for me.
- I perceive a threat potential from not knowing this.
- I felt it was not relevant for my decision of where exactly to start.
Result and
Discussion
Remarkable is that it was perceived as unknowable by the interviewee in the decision
situation. Experienced scientists (experts) say this would be clearly defined. It might
be that some additional intellectual capacity is needed that can be only developed
over time (experience) or by being very reductionist and following cookbook style
instructions of what scientific quality requirements are.
Conclusion
It could be concluded that the ambiguity was so high for the decision maker that
something knowable by experts was not perceived as knowable. Hence individual
capacities
to reduce ambiguity were the main impediment of knowledge. Not
information.
Map a “value” related aspect of non-knowledge from the interviews from the perspective of the
interviewee “if my proposal will be better” – an if to decision.
Reflections
on
Dimensions
- Recognised.
- I think it is knowable because my boss could know (better).
- It relates to the future.
- I also assume the reaction, so if my proposal will be better, to be ambiguous.
- For my solution it is relevant if my proposal will be better.
- It is unintentional that I do not know if my proposal will be better.
- If my proposal will be better is potentially benefiting.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
127
Result and
Discussion
It is remarkable how the interviewee defines knowable – by a better approximation
which the boss is supposed to have. But the boss was not asked because of limiting
circumstances. Is this really unintentional? In this understanding, knowable means
trained by experience or knowable for someone with different intellectual capacities.
It could also be now unknowable –
as it will be only known once others react to the
new proposal. Then it would also not be relevant to the solution, as the solution to
write a new proposal has to be taken, by definition, without knowing if the proposal
is better. The proposal is potentially
benefiting because its current alternative is
dissatisfying, the interviewee hence perceived the need to improve the proposal. The
interviewee probably has some quality standards in mind that might not be equal to
the quality standards the boss has. But the boss´s quality standards are not accessible
at the moment. It can be asked where to stop the run for optimization. The
interviewee might have used a stopping rule similar to this: “I propose something and
I know it could be improved by my boss but if my boss is not available my version will
do”.
Conclusion
Stopping rules are effective and necessary heuristics to decide in the unknown.
Knowability can be defined in relation of the own to someone else´s knowledge
Map an “others” related aspect of non-knowledge from the interviews from the perspective of the
interviewee “intentions of the government” – a how to decision.
Reflections
on
Dimensions
- Recognised.
- It extends through the present.
- It is unknowable
as it is the institutionalised collective mind of various
individuals in a power structure and I am from a different culture and my
approximation abilities are limited.
- It is ambiguous
. I assume not everyone involved in the government has the
same intentions and it is not determined how their collective intentions will
materialise for different environments and questions.
- It is potentially benefiting.
- It is unintentional.
- It is irrelevant for my solution because I will propose the technically best
options regardless of what the government wants. My decision is just a
proposal so we will deal with the government´s behaviour in later decisions.
Result and
Discussion
The interviewee anticipated relevant non-knowledge for a later step in the process.
When asked, the interviewee responded this non-knowledge would not be relevant
for the concrete decision point. It can be discussed if such an early anticipation is
unnecessarily distracting or
a good thing to be kept in mind during preceding
decisions. Ambiguity is the main dimension which limits the knowability. Knowability
is also limited by complexity.
Conclusion
Irrelevant non-knowledge might require redefining the decision point analysed.
Map a known “the sky is blue”
Reflections
on
Dimensions
- Unintentional (innate perception + linguistic agreement).
- Unambiguous (unless for blind people).
- Wide social distribution (except the blind).
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
128
- Manifest in perception and linguistic concept.
- Potentially benefiting.
- Recognised.
- Knowable.
- It extends through the present.
- Solution relevant?
Conclusion
Knowns can be mapped with the non-knowledge map.
The following example was generated with the updated and board game non-knowledge map (see
annex VII.1). The user was left to read the rules of the game but could ask the author.
Map “if results will improve significantly” for the if to decision “should I go on with model
calibration”.
Reflections
on
Dimensions
- Extends through the present.
- Ambiguous, but the term is unclear to me.
- Unknowable.
- Relevant.
- High threat potential, but it could also be potentially benefiting (if results
improve significantly)
- Unintentional.
- Recognised.
Tools (proposed elements and principles of non-knowledge literacy)
I always use the joker (delay the decision)
Dismiss the model: No! (
Change to a solution where the
unknown does not matter)
- Stopping rules is what I would need!
Result and
Discussion
The user had difficulties to define the decision point and the unknown and changed
both iteratively. The user also brainstormed on the properties of the unknown but
results remained rather one-dimensional speculations about knowability (hence not
displayed). The user said potentially benefiting and high threat potential should not
be part of the same dimension, as it was difficult to decide between both. In
discussion, the user decided to put threat potential, as the fear seemed larger than
the hope. The user had oscillated during various months between the options to
continue calibrating the model as the expected gains were high and the current
results were not satisfying and the threat potential that any next try would cost an
unknown lot of time while delivering results that would be an unknown lot better.
The term ambiguity remained unclear to the user; reflections with the author
included an estimate of the risk.
The formulation of the unknown is problematic, as it embraces other unknowns
(significant is not defined). It might help to map “what improvement will be
significant”. Displayed mathematically, the problem would be a curve between time
(effort) and advantage, the course of the curve is unknown and it is unknown if it will
ap
proach an asymptote or if jumps can be expected. However, the threat potential
dimension was not related to the unknown, as the unknown is formulated potentially
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
129
benefiting. As the unknown (something like the marginal utility) is unknowable, a
heuristic is needed to stop searching. Gigerenzer proposes various stopping rules and
everyone can invent one.
Conclusion
Formulation of unknown and decision are crucial. Ambiguity is a difficult term. There
are unknowns which essentially are non-knowledge if somethi
ng has a threat
potential or is potentially benefiting. Then this dimension cannot be sensibly
answered.
Characteristics, relevance and dynamics of the unknown are suggested as starting points (Ibisch
and Hobson, 2014)
Reflections: The characteristics of the unknown would be perceived by an individual or a group and
sometimes in comparison to world knowledge. This would represent the manifestation of the
unknown and the related known. Characteristics could be the dimensions. The relevance is related to
a specific decision point and gets more ambiguous if applied to processes or open problems. The
dynamics of the unknown require a perspective over time. This can be done by the graphical
representation of the map which maps (non-)knowledge against time.
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
130
VIII. Digital Annex B
See pdf and xlsx documents on the attached compact disk. The numbered items in the list below are
bookmarked in the pdf document.
1. Non-knowledge Literacy Quotes
2. Questionnaire Versions 0.0-1.4
3. Terms for Coding (how did you decide)
4. Clustering for Coding (how did you decide)
5. PCA Results
6. PCA Data
7. Coding (Keys and Abbreviations used for PCA)
8. Clustering (Mapped) (Non-)knowledge
9. Exposé
10. Development of Questionnaire
Lara Mia Herrmann – Master Thesis The Non-knowledge Map for Decisions
131
IX. Declaration of Independent Work on Master Thesis
With this statement I, Lara Mia Herrmann, declare that this Master thesis “The Non-knowledge
Map for Decisions – A Tool to Explore and Create Non-knowledge Literacy” was prepared by
me, only using the given references in this paper.
Berlin, 13th November 2015
