Futures literacy: A hybrid strategic scenario method

Abstract

What stories do we tell about the future? This article develops a topology of storytelling about the future, which is used to develop a definition of ‘futures literacy’. It goes on to outline a hybrid strategic scenario method for acquiring the capacities of futures literacy.

Full text

This article was originally published in a journal published by
Elsevier, and the attached copy is provided by Elsevier for the
author’s benefit and for the benefit of the author’s institution, for
non-commercial research and educational use including without
limitation use in instruction at your institution, sending it to specific
colleagues that you know, and providing a copy to your institution’s
administrator.
All other uses, reproduction and distribution, including without
limitation commercial reprints, selling or licensing copies or access,
or posting on open internet sites, your personal or institution’s
website or repository, are prohibited. For exceptions, permission
may be sought for such use through Elsevier’s permissions site at:
http://www.elsevier.com/locate/permissionusematerial

Author's personal copy
FUTURES
Futures 39 (2007) 341–362
Futures literacy: A hybrid strategic scenario method
Riel Miller
!
Senior Visiting Fellow, Danish Technological Institute, Tecknologiparken Kongsvang Alle 29,
KD-8000 Arhus C, Denmark
Abstract
What stories do we tell about the future? This article develops a topology of storytelling about the
future, which is used to develop a definition of ‘futures literacy’. It goes on to outline a hybrid
strategic scenario method for acquiring the capacities of futures literacy.
r2006 Elsevier Ltd. All rights reserved.
The human condition can almost be summed up in the observation that, whereas all
experiences are of the past, all decisions are about the future. The image of the future,
therefore, is the key to all choice-oriented behavior. The character and quality of the
images of the future which prevail in a society are therefore the most important clue
to its overall dynamics.
Kenneth Boulding, Foreword in The Image of the Future, Fred Polak, 1973.
0. Introduction
People think about the future all the time. In the morning, when they wake up and start
planning for the day ahead. At the dinner table, when they discuss how to pay the bills next
month or what will happen in the Middle East. Most of these reflections are about the
short term, a few days, weeks or months. Such conversations are a natural mix of people’s
hopes and fears with a wide range of probable to improbable expectations. Professional
forecasters trying to predict tomorrow’s weather or next year’s economic growth handle
degrees of probability more carefully. Professionals tend to focus on getting to the highest
ARTICLE IN PRESS
www.elsevier.com/locate/futures
0016-3287/$ - see front matter r2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.futures.2006.12.001
!
Tel.: +336 3239 1360.
E-mail address: rielm@yahoo.com.

Author's personal copy
probability prediction that available models and data can provide. They generally steer
away from considering the broader, less predictive question of what might be possible as
well as the more normative question of what is desirable. This choice, while under-
standable in light of our deep hunger to know the future, has important limitations.
Two drawbacks of the search for greater predictive accuracy are worth highlighting.
First, there is the familiar risk of adopting forecasting methods and models that depend
too heavily on what happened in the past. Yesterday’s parameters may do a good job at
tracking past events, but experience shows that this approach consistently misses major
inflection points related to long-run transformative changes [1–5]. A second, less remarked
danger is that a preoccupation with what is likely to happen tends to obscure outcomes
that may be unlikely but still possible and potentially more desirable. At best, seeking the
reassurance of greater predictive certainty tends to overlook what is viewed as less likely.
At worst, it lulls us into a false sense of having exhausted the available options, thereby
narrowing the set of choices considered actionable. This, in turn, can impair strategic
decision-making
1
because it leads to an underinvestment in imagining less predictable
outcomes and the avenues for reaching them.
Futurists, in their efforts to satisfy the yearning for ‘‘scientific’’
2
consideration of the
future, have been aware of these problems for a long time [6,46,47,52,53]. Currently, most
futurists formally recognise the impossibility of assigning meaningful probabilities to the
way a society or organisation might function in 10 or 20 years. Notwithstanding this
formal nod to the inadequacy and failures of the predictive quest, little progress has been
made in developing practical alternatives [54]. Three of the many reasons for this
inconsistency between theory and practice were/are: a fear of the future that drives a deep
desire to know (divine) what will happen (clients want predictions); recent (post WWII)
relative systemic stability and the related success of planning in this context
3
; and lack of
experience with, and hence under-development of, the conceptual tools and behavioural
conventions that make it practical to embrace non-predictive approaches to decision-
making with success. Old paradigms do not cede easily, and the attachment to predictive
approaches rooted in trend analysis, forecasting models, multi-factor calculations, etc.,
is tightly integrated with the way risk is managed and decisions are taken in industrial
society.
The premise of this paper, without claiming any degree of probability, is that the world
around us today, in its evolving conceptual and practical attributes, is creating a context
that, on the one hand, is dispensing with industrial era modes of perpetuating systemic
stability, risk management, decision-making, etc., and, on the other hand, is embracing
complexity, heterogeneity and spontaneity as opposed to simplification, homogeneity and
ARTICLE IN PRESS
1
Strategic is used here in the sense of an open (i.e. un-predetermined) selection of an objective(s) (what matters)
and, subsequently, an open search for the choices deemed essential for realising the goal(s) (what matters for what
matters).
2
Scientific is defined here in a rather minimalist fashion as analysis based on explicit and open methods that test
hypotheses pertaining to a particular subject through inter-subjective evaluation.
3
This point is a major topic in its own right. Briefly, the main hypothesis is that the post-World War Two social
order, in most industrial countries, was dominated by closed hierarchical decision-making systems that generated
bounded predictive planning horizons consistent with the (partly self-fulfilling) continuity of continuous
improvement to that system (Kaizan). Two distinctions are important here, one between transformative and non-
transformative incremental change, and the other between systems where planning (ex-ante choice) sustains the
system and those where it does not.
R. Miller / Futures 39 (2007) 341–362342

Author's personal copy
planning.
4
As a result, it is hypothesised that modal or non-predictive narratives
(scenarios) [7,8] could become more important and more practical, for two reasons.
One reason is that if the conditions for change in systemic sustainability (change in the
conditions of change [9]) are emerging, then the role of ‘‘what-if’’ imagining becomes
central to making the choices (individual and collective) that, in turn, are pre-requisites for
actively pursuing such ‘‘transition scale’’ change [10]. In other words, explicit choices in
both the individual (behaviour, aspirations, etc.) and collective (institutions, rules, etc.)
spheres are fundamental to operationalising societal transformation and, hence, so are the
methods for imagining the choices that are outside the old order.
5
Second, more specific to this particular historical period is the growing importance of
non-predictive imagining owing to a shift towards greater spontaneity in decision-making.
Learning-by-doing, experimentation’s successes and failures, is generating the concepts
and techniques for embracing the richer, more accurate information available to just-in-
time decision making.
6
Practice is evolving the networking, diversification, openness, trust,
and fluidity that makes it feasible to be spontaneous and rely on complexity rather than
fight it.
7
All of this calls for an enhanced capacity to imagine the possibilities of the
moment without succumbing to the temptation to plan the future based on probabilistic
calculations about the unknowable [8].
8
Instead, as discussed in more detail at the end of
this article, the capacity for more imaginative storytelling becomes the way to use the
potential of the present more effectively and, assuming that people’s choices are consistent
with their values, evolving in ways that realise our aspirations.
Clearly the hypothesis that non-predictive scenarios are gaining in importance is
itself rooted in a scenario of how post-industrial society might function [10–13], a reminder
that there is no escaping the interpretive lenses that always filter, reflexively, our
perceptions and analyses. From this self-awareness starting point, the rest of this
article offers a partial overview of one way of making sense of how stories about the
future may be developed and used in our emergent, perhaps eventually post-industrial
societies.
ARTICLE IN PRESS
4
Systemic change of this kind is very rarely absolute. Rather, it is compositional in nature, i.e., the shares of
certain types of activities in the total changes. Most often, such changes in overall composition occur owing to the
ascendance of new components that gradually move toward preponderance rather than a decline (in absolute
terms) of the old components. In other words, in certain domains, the simplification, homogeneity and planning of
mass-society continues, but as a much smaller relative share of the total activity of society.
5
NB: Revealing choices does not necessarily imply anything about the probability of particular outcomes.
6
Where P
d
is a planned decision based on the information available prior to the decision I(t!1), S
d
is a
spontaneous decision based on the information available both prior to and at the time of the decision (I(t!1)+I
t
),
therefore, all other things being equal, there is more information available for S
d
than P
d
. In part, this article is an
analysis of one of the conditions that would need to hold for this assumption to be true, specifically the capacity to
ensure that I
t
!1 does not reduce I
t
but enhances it.
7
Changes of this kind are often generational, meaning that young people who swim in the waters of complexity
are not only more at ease than their elders who grew up on land, a less complex and mostly two-dimensional
terrain, but are capable of conceiving and doing things that the elders simply cannot. Such generational shifts that
transform, for instance the perception and management of risk, may be one of the primary attributes of changes in
the conditions of change of social systems.
8
The border line between possible and probable is not a fixed line. Both are estimations of the likelihood of an
event actually happening in the future. However, possibility slips only modestly over the threshold from
impossible to plausible while probability pushes above this minimum to become a positive degree of predicted
occurrence (a point discussed further in Section 2.2).
R. Miller / Futures 39 (2007) 341–362 343

Author's personal copy
1. The challenge of choosing a story to tell
Stories may be what make life intelligible. Certainly, humans invent and tell many kinds
of stories for many different purposes. Future scenarios (FS) are just one specialised type
of story, most often invoked in the context of complex and ambiguous situations [14–17].
As with any story, one of the primary challenges is how to imagine and then select a few
distinctive and pertinent narratives from the vastness of the imaginable [18]. There are two
commonly used methods for solving the problem of how to choose specific stories about
the future.
The first, springing directly from the predictive tradition, takes an initial starting point,
for instance population or economic output, and then develops scenarios on the basis of a
range of growth rates—low, medium and high. This method can be called the baby-bear,
momma-bear and papa-bear approach (Bear for short).
9
The second technique focuses
more on preferences and expectations in order to sketch scenarios that capture the stories
of the futures that people consider to be: the most desirable, the least desirable and the
muddling through version that mixes a bit of good and a bit of bad (which also usually
happens to be the one people consider most likely or realistic). This method can be dubbed
‘‘the good, the bad and the ugly’’ approach (GBU for short).
Both of these methods generate FS with the virtue that the stories are usually quite
familiar. For instance, it is easy to recognise Bear scenarios regarding the future of, for
instance, universities that are distinguished by differences in enrolment growth rates—low
(baby-bear), medium (momma-bear) and high (papa-bear). Or GBU scenarios that are
distinguished by the preferences of people whose values, for instance, lead them to consider
the ‘‘good’’ scenario to be one where universities are exclusively citadels of a pure search
for knowledge, the ‘‘bad’’ scenario to be one where universities are exclusively driven by
the commercial imperatives of funders from the private sector, and a muddling through or
‘‘ugly’’ scenario, usually seen as the most likely, that combines both pure and commercial
options.
Many foresight exercises generate a composite scenario matrix by mixing and matching
a variety of trends (Bear) and preferences (GBU). For example, Table 1 combines three
sets of scenarios (1 GBU and 2 Bear) for the future of tertiary education. Purely for
the sake of illustration, the GBU scenarios, based on knowledge-driven versus
commercially-driven models of tertiary education, are crossed with two categories of
Bear scenarios based on (1) technological change and (2) enrolment growth rates. Building
up this kind of matrix can rapidly generate numerous stories, 18 in this case, for the future
of tertiary education.
Relying on Bear/GBU methods to produce FS has its strengths and weaknesses. On the
positive side, Bear scenarios can lean heavily on readily identified trends and the findings of
the ‘‘predictive social sciences’’ such as economics and demography. Whereas GBU
scenarios, generally developed by teasing out the expectations and values of specific
constituencies like business executives, groups of experts, citizens in a local community,
ARTICLE IN PRESS
9
Predictive scenarios can, and regularly do, use sophisticated modelling techniques, sometimes originating in
theories of change, that are specified using sets of inter-dependent and multivariate, often non-linear equations.
Short-run economic forecasts are a pre-eminent example. However, these scenarios remain predictive in both
intent and conception, and are most often developed and tested by fitting the model’s parameters to past data in
the hope that a similarly good fit will occur in the future [19]. Applying models to the past as a way to ‘‘test’’ a
hypothesis is how social science usually tries to mimic the experimental methods of ‘‘hard’’ or ‘‘natural’’ science.
R. Miller / Futures 39 (2007) 341–362344

Author's personal copy
etc., usually have the virtue of generating stories that are, at least initially, easy to
comprehend.
On the negative side, there is an obvious problem when the scenario matrix gets out of
hand as more variables, rates of change and preferences are added. The explosion of
permutations and combinations reflects the richness of people’s perceptions, but makes it
difficult to distinguish the different stories in meaningful ways.
10
Even more damaging, as
anyone who has experienced these processes knows, because the variables are often highly
incommensurate and lack explicit and/or tested hypotheses regarding basic definitions,
causal inter-relationships or weighting in terms of the model’s parameters, it becomes
difficult to use these scenarios for deeper analysis and decision-making [21].
How serious are the limitations to FS arising from a reliance on Bear/GBU methods as a
way to select and specify stories about the future? The answer to this question depends
almost entirely on the purpose or task to which the scenarios are being put. There are
clearly certain tasks, such as building temporal awareness and revealing shared or
divergent assumptions and goals, where using GBU/Bear methods is highly effective,
precisely because it surfaces values and expectations. However, crucially, when it comes to
policy analysis and strategic decision-making, the flaws in these two approaches to
building scenarios are serious enough to compromise effectiveness.
Three problems with Bear/GBU methods for generating analytical and strategic FS
come to the forefront.
First is the limited realm of the possible when, as in most cases of applied scenario
development, the GBU/Bear narratives originate either in people’s current (usually
unexamined) expectations and preferences or in the framework of predictive modelling
ARTICLE IN PRESS
Table 1
Mixing Bear and GBU scenarios
Bear scenarios The Good, The Bad and The Ugly
Knowledge driven Commercially driven Mixed model
Low rate of tech change
Low enrolment growth Scenario 1 Scenario 2 Scenario 3
Medium enrolment growth Scenario 4 Scenario 5 Scenario 6
High enrolment growth Scenario 7 Scenario 8 Scenario 9
High rate of tech change
Low enrolment growth Scenario 10 Scenario 11 Scenario 12
Medium enrolment growth Scenario 13 Scenario 14 Scenario 15
High enrolment growth Scenario 16 Scenario 17 Scenario 18
10
Techniques like cross-impact matrices can be helpful as a way to winnow out the permutations into a handful
of scenarios that can be analysed [5]. However, unless a careful distinction is made (see Godet [5]) between
participatory and analytical phases, there is a significant danger of constructing models where the variables and
hypotheses that make up the model are impressionistic rather than analytical, even when based on ‘‘experts
opinions’’, leading to the old problem that even if the cross-impact calculations are done using sophisticated
techniques, it is still a problem of ‘‘garbage in-garbage out’’. If the reduction process does not follow rigorous
selection criteria that reflect a theoretically coherent model to reduce the number of outcomes, then not only are
there arbitrary losses of information, but there is also a serious risk of poor specification robustness of the
outputs, i.e., the reduced set of scenarios [20].
R. Miller / Futures 39 (2007) 341–362 345

Author's personal copy
familiar in the social sciences. As a result, non-conventional and transformative
possibilities are under-represented.
11
Second, as already noted, it is more difficult to push
the frontiers of the imaginable in actionable (policy choice) ways when the methods used
for specifying the scenario variables, usually a structured conversation among pertinent
constituencies, are rarely rooted in explicit, rigorous
12
analytical-systemic frameworks.
Third, the virtue of distinguishing scenarios using either values or expectations turns
into a vice when the objective is to imagine a set of equally preferable (isovalue) or equally
probable (isoprobable) outcomes or paths for achieving the same outcome. GBU and Bear
scenarios, within respective spheres, are inherently different, not equal, when it comes to
values and expectations. That is what distinguishes one scenario from another. But, as a
result, it is more difficult to generate equally desirable scenarios (isovalue) that can be used
to analyse different ways of getting to similar goals or equally likely scenarios
(isoprobable) that can facilitate choosing from amongst outcomes that reflect different
values. This drawback of FS generated using GBU and Bear methods is a significant
handicap for decision makers who are often called upon to make choices between equally
desirable or equally (un)likely futures.
To be clear, the weaknesses identified here are neither due to the nature of these
scenarios as evocative stories about the future nor the utility of using expectations and/or
preferences to develop such narratives. The problem is that an exclusive reliance on
predictive and/or value-based FS results in an incomplete and, hence, inadequate
interrogation of the potential of the present. Evidence of this poverty of invention, narrow
repetitiveness of findings and modest utility, in terms of an impact on strategic decision-
making, can be found, for instance, in much of the scenario work done in the education
sector, be it on the future of universities [26] or of schooling [27,28]. The scenarios that
predominate are generally the product of single iteration
13
GBU and Bear exercises,
ARTICLE IN PRESS
11
The epistemological divide between the predictive and non-predictive methods has been explored in the debate
about modal and non-modal statements [7,22,23]. Also, Karl Popper in his critique of historicism points to the
importance of grasping the possibility of ‘‘changes in the conditions of change’’ [9]. While Douglas North points
out that conventional social science does a poor job with the non-ergodic [24]. Kenneth Boulding points out that
evolutionary systems have changing parameters and are, therefore, impossible to predict [25]. Ilya Prigogine
argues that ‘‘the future is no longer determined by the presenty. Mankind is at a turning point, the beginning of a
new rationality in which science is no longer identified with certitude and probability with ignorance’’ [8].
12
Use of the term ‘‘rigour’’ may seem value-laden, but the intention here is simply to distinguish between stories
based on explicit theories of a system (static or dynamic) that are the bread and butter of social science (hypothesis
testing) and stories that do not spell out the underlying models relying instead on ad hoc impressions, untheorised
expectations, or, in the case of views pooled from ‘‘experts’’ (often called a Delphi exercise), a grab bag of theories
that have not been tested for commensurability or compatibility. I am not arguing that the former will be more
accurate in ‘‘predicting the future’’ than the latter. In my view, neither is adequate to that task because long-run
prediction of social systems is not only impossible, but a breach of the contingency principle that is fundamental
to notions of ‘‘free will’’ and ‘‘creating the future’’. Rather, the distinction between rigorous and non-rigorous
scenarios is the extent to which the variables and hypotheses that constitute the story have been explicitly specified
in ways that respect basic rules of coherence [20]. GBU/Bear scenarios can be generated in a rigorous manner;
only as explained below, it is usually an unnecessary investment given the aims of most such scenario exercises.
13
Two different aspects of the distinction between single and multiple iteration scenarios are pertinent here. The
first, more substantive distinction refers to the contrast between sets of scenarios derived, through more or less
intricate processes, using one basic methodology and other scenarios that are produced by generating a series of
scenario sets with each set developed using a different method. The ‘‘hybrid strategic scenario’’ (HSS) method
explained below is just one example of a multi-step, multi-method scenario development process. The second, less
important dimension has to do with the number of iterations of the processes associated with a single scenario
method.
R. Miller / Futures 39 (2007) 341–362346

Author's personal copy
sometimes mixed together to produce composite scenario matrices [29]. Similar drawbacks
have been noted with respect to many technology foresight initiatives [30]. Without calling
into question the positive outcomes of these efforts for teamwork and leadership, this type
of exercise has generally failed to produce analytically persuasive or actionable conclusions
for policy, particularly policies intended to take into account transformative societal
change [19].
There is another, less remarked danger that arises from an uncritical reliance on GBU/
Bear methods for selecting stories about the future: it is the difficulty of determining which
methods are most appropriate for particular tasks. Existing typologies of scenario
activities are largely inductive, arriving at the categories by reviewing existing practices
differentiated on the basis of the scenario process or goal [18]. A more deductive
approach, moving from general assumptions like the modal attributes of a futures studies
methodology I have called ‘‘history of the future’’ [31], can generate a capacity-related
typology that treats storytelling about the future as a capability or capacity that can be
more or less sophisticated. Taking the capacity point-of-view for constructing a typology
gives rise to what I call a ‘‘Futures literacy’’ (FL)
14
framework for categorising thinking
about the future.
2. A hybrid strategic scenario method for developing FL
FL is the capacity to explore the potential of the present to give rise to the future. Like
language literacy, FL is a variable or cumulative capacity that can be used for many
purposes, ‘‘good’’ and ‘‘bad’’. One level is not ‘‘better’’ than another, but Level 1 is a
precondition for Level 2, and Level 2 is a precondition for Level 3. This is no different than
reading. Learning the alphabet is the first step in acquiring the capacity to decipher text,
and then being able to read is crucial for beginning to decipher the messages contained in a
text. Again, like reading, FL is not the same as the text that is being read or producing a
text to be read. Certainly, the capacity to read a text, in its many senses, can be improved
through practice, including writing text. But FL is not ‘‘the future’’; it is the capacity to
think about the potential of the present to give rise to the future by developing and
interpreting stories about possible, probable and desirable futures.
One way of learning and practicing FL is to take a hybrid and sequential approach. The
hybrid method is used because the only way to ensure congruency between scenario
techniques and tasks is to use a range of different approaches. The sequential method is
used because getting to Level 3 FL requires the capacities and knowledge acquired at
Levels 1 and 2. The hybrid strategic scenario (HSS) method developed here, in keeping
with the cumulative nature of FL, involves working through and acquiring the skills
associated with all three levels of FL as summarised in Table 2.
2.1. Level 1 FL: awareness
Level 1 FL is largely about developing temporal and situational awareness—meaning a
greater appreciation that change happens over time and that particular constituencies,
products or organisations can be situated in time according to their values and
expectations [14–16]. By revealing common goals and shared assumptions, the typically
ARTICLE IN PRESS
14
The term ‘‘futures literacy’’ has deep roots in futures studies as a field, see for instance [6,32].
R. Miller / Futures 39 (2007) 341–362 347

Author's personal copy
discursive group processes used to develop Level 1 FL help forge stronger teams and build
the kind of confidence needed to make decisions about the future. Much applied foresight
is practiced at this level and has an admirably successful track record at improving team
and leadership capacities [21,27]. Bear/GBU scenarios, and particularly the learning
processes used to generate them, are highly effective at developing the awareness that
change happens over time, that people do harbour expectations and values, and that
choices might matter.
Techniques for building Level 1 FL are legion and, in most cases, do not require a very
elaborate catalyst to get the process going.
15
People are primed to express their views about
the kind of future they prefer and expect. However, the ease with which dinner table or
workshop conversations carry through to the construction of more elaborate scenarios can
be deceptive. While it is one thing to discuss expectations and preferences, it is another to
weave them into a convincing story. Equally difficult, as teachers know, is to engage in a
learning process that develops and embeds new capacities. Meeting both of these
challenges calls for careful elaboration of the scenario process, following the rules of good
pedagogy and storytelling. Leaving aside the issue, dealt with extensively in many different
fields including future studies [32] of how to design an effective learning process, the issue
of narrative structure or the rules of storytelling when it comes to FS merits a brief
ARTICLE IN PRESS
Table 2
Levels of futures literacy—tasks and techniques
Futures literacy Task Technique(s)
Level 1
awareness
Temporal awareness, shifting both values
and expectations from tacit to explicit—all
of which builds the capacity of people, teams
and leaders to respond and innovate
A wide range of catalysts and processes
generate the discussions and sharing of
stories that elicit people’s views on what they
want and expect in the future
Level 2
discovery
Rigorous Imagining (RI) involves two
distinct challenges—imagination and rigour,
the former in order to push the boundaries
and the latter so that what is imagined is
‘‘scientific’’
a
and intelligible
Escaping from the probable and preferable
to imagine the possible demands systematic
creativity and creating systematically, non-
discursive reflection and social science are
essential ingredients
Level 3 choice Strategic scenarios are aimed at questioning
the assumptions used to make decisions in
the present, not as targets to plan-by but to
provide new insights into the potential of the
current world as a way to embrace
complexity, heterogeneity and the pertinence
of spontaneous actions that put values into
practice
Strategic scenarios are constructed using the
capacities and stories acquired in developing
Levels 1 and 2 FL, by combining values,
expectations and possibilities into scenarios
that follow the narrative rules (see Level 1
FL below) and the methods of ‘‘history of
the future’’ [31]
a
See footnotes 2 and 12.
15
Big investments in preparing data about trends or generating robust scenarios as catalysts for these Level 1
participatory processes are generally a waste of resources. People want to dive into the discussion of their
expectations and preferences, and huge prior documentation just gets in the way even as it reassures nervous
‘‘experts’’ who often implicitly realise that the predictive and analytical content of Level 1 exercises is very low.
They invest in ‘‘evidence’’ in the vain hope that this will lend ‘‘scientific’’ weight to what in the end can only be,
given the techniques used, an impressionistic, dialogue building exercise. Excessive investments of this kind reflect
not only a failure to correctly match methods to tasks but also a fundamentally non-modal view of the future.
R. Miller / Futures 39 (2007) 341–362348

Author's personal copy
discussion. At a minimum because it is an aspect of scenario works that is often
overlooked.
The narrative of FS, like familiar ‘‘artistic fiction’’, is more intelligible and, hopefully,
meaningful if it follows certain rules. In particular, it is important to be clear about the
following five attributes of the story:
(i) What is the type or purpose of the story? Not tragedy or comedy, thriller or romance;
but is it for contingency planning, simulation training, team building, optimisation
testing or strategic imagining?
(ii) What is the temporal or chronological frame? Not beginning, middle and end; but
comparative static (two or more cross-sections), dynamic trajectory (time-series) or
back-casting (reverse engineered)? And, in practice, what is the time-span under
consideration (years, decades, generations)?
16
(iii) What is the (analytical) point-of-view? Not first or third person, stream-of-
consciousness or dialogue; but is the story told in terms of the choices people make
in their everyday lives (micro), or does it describe aggregate outcomes (macro) or is it
about both (linking micro and macro)?
(iv) What are the main protagonists (ultimate decision makers)?
17
Not hero and villain;
but is the agent using the scenario to assist with making choices at a specific institution
(firm, school, hospital, etc.) or about a social/economic system (nation, sector, etc.)?
And what are the relationships between the agents as decision makers? Not parent/
child or detective/criminal; but is the story of a specific agent like a school or a
financial market stand-alone, told in terms of its internal functioning only, or is it
embedded and interacting with a story of surrounding societal changes in production,
consumption, uses of technology, social status, codes of conduct, etc.?
(v) What rules apply to the action? Not is time travel allowed or Matrix (the film) like
suspension of the rules of this Universe’s physics; but what are the assumptions and
hypotheses that underpin the analytical definitions and causal relationships that make
for robust social science?
Explicit answers to these five questions go a long way towards ensuring that scenarios
tell an internally consistent story. Helping both the teller and listener to grasp how time,
values and expectations shape our understanding of the potential of the present.
The easiest way to clarify what distinguishes the different levels may be to provide an
illustrative example using a what-if exercise. In this case the what-if exercise is: what if
there had been an HSS process undertaken in the late 19th century with the aim of thinking
about the future of the new and exciting technology of electricity. Imagine that it is the late
19th century, that electricity is in its infancy and that a group of futures oriented thinkers
from the private and public sectors get together to assess the prospects for this new
technology. As they discuss its potential, they realise that the stories they are telling are
very open ended, not only because it is difficult to envisage how the technology will be
ARTICLE IN PRESS
16
From a strictly logical point of view, everything that is not the present is unknowable, and in so far as FS are
not about prediction but imagination, the actual time-frame is beside the point. However, the time-frame does
have an impact on the fifth point about the rules of action. Usually, over longer time periods, more parameters
become open, thereby reducing the number and potentially the importance of the ‘‘framing’’ assumptions.
17
Often the protagonist, naturally in the ‘‘starring role’’, is the client or interested party that sponsored or
undertook the HSS process.
R. Miller / Futures 39 (2007) 341–362 349

Author's personal copy
used, but also because industrialisation and globalisation are in full swing. As they refine
the stories about the future of electricity, they share specific values about access to
technology, the role of technology in society and the political principles they believe should
underpin decision-making—who and how to decide.
In the end, they produce a set of GBU scenarios. In the ‘‘good’’ scenario, electricity is
used to build much bigger factories than was possible with steam or water power and to
illuminate the bigger cities that accompany bigger factories. In the ‘‘bad’’ scenario,
electricity turns out to be too expensive and dangerous to play a really major role, while in
the muddling through ‘‘ugly’’ scenario, a bit of both the good and the bad lead to a
reduced but still important role of electricity in the ‘‘factory era’’. Of course, it would have
been very difficult to anticipate that electric motors would become so small and mobile or
that power generation could be so effectively centralised and distributed. The predictive
capacity of these fictitious futurists is, as expected from a modal perspective, very low.
However, in terms of Level 1 FL goals, the process has achieved its key targets. The group
has rendered explicit a number of values and expectations. They have discovered elements
of a ‘‘common language’’ for talking about change, and they have probably made
considerable progress in defining the subject—the potential role of electricity in society—
more clearly. However, even the best Level 1 FL stories suffer from the limitations
discussed previously. Level 2 FL or ‘‘rigorous imagining’’ is one approach to overcoming
the pitfalls of Level 1 FL.
2.2. Level 2 FL: discovery
Level 2 FL is the capacity to overcome the limitations imposed by values and
expectations when thinking about the future. It is a technique for conducting the
potentially paradoxical task of ‘‘rigorous imagining’’. This is a crucial and challenging step
in opening up new insights into the nature and determinants of today’s potential. Rigorous
imagining depends on carefully and consistently distinguishing between possible, probable
and preferable. Such distinctions are necessary for rigorous imagining because the task of
imagining possible futures is logically and practically prior to the assessment of
probabilities and preferences. It is prior to the assessment from a logical perspective
because, as illustrated in Fig. 1, preferable and probable futures are subsets of the possible,
and prior to the assessment from a practical perspective because as already pointed out,
consideration of preferences and probabilities constrain the imagining of possibilities.
18
Starting with the logical, a useful first step in expanding the range and analytical content
of possible futures is to clarify, at a conceptual level, what distinguishes conceivable,
possible, probable and desirable scenarios. Fig. 1 is an illustration of one way of defining
these categories and the logical inter-relations. The largest set consists of conceivable
scenarios (Zone 0). A sub-set of conceivable futures are those that are possible (Zone 1).
Nested within the set of possibilities are probable futures (Zone 2) and some of
the desirable ones (Zone 3).
19
There are, of course, desirable futures that are neither
ARTICLE IN PRESS
18
Note that this does not mean skipping Level 1 FL because a basic understanding of contingency and an
awareness of current assumptions (values and expectations) is crucial for overall FL. However, Level 2 FL does
require a break with Level 1 FL because the point is to at least make explicit current assumptions if not to begin to
discover what it means to develop alternatives.
19
As desirability is in the eye of the beholder, this set contains both good and bad scenarios; it depends on your
point of view.
R. Miller / Futures 39 (2007) 341–362350

Author's personal copy
probable nor possible, even if conceivable. Hence Zone 3 spills over from Zone 2 into
Zones 1 and 0.
20
Describing, as in Fig. 1, the infinite number of stories that make up the whole scenario
space, does not in-and-of-itself solve the problem of how to choose specific stories from
within that space. As already discussed, GBU/Bear methods offer one way of identifying
the variables and selecting scenarios (Zone 4), but in ways that are limited by the pursuit of
scenarios based on values and/or probabilities. The challenge then is, in most cases, to use
non-predictive, value neutral methods to identify both the variables and the scenarios that
are possible (Zone 5).
Rigorous imagining can be conducted in a multitude of different ways. However, it is
one of the distinguishing features of Level 2 FL that many of its story selection methods
are incompatible with those of Level 1 FL. The definitional differences are obvious as
discovering what is possible (minimally probable) is not the same as determining what is
likely (more than minimally probable) or desirable. The practical differences arise from the
nature of the competences entailed by different levels of FL. Level 1 FL largely involves
shifting knowledge from its tacit to explicit form, i.e., what people already know about
time, preferences and expectations. The pedagogy for achieving Level 1 FL calls for
structured learning processes that help reveal to people their existing assumptions.
ARTICLE IN PRESS
Conceivable
Futures (Zone 0)
GBU &
Bear
Scenarios
(Zone 4)
Possibility
space
stories
(Zone 5)
Strategic
Strategic
Strategic
Probable
Futures
(Zone 2)
Possible
Futures
(Zone 1)
Desirable
Futures
(Zone 3)
Fig. 1. Strategic scenarios and possibility space scenarios.
20
There are possible and probable scenarios that are not conceivable. However, using the criterion of
practicality, I have limited the conceptual space to those that are conceivable [23]. Furthermore, I have nested
desirable mostly within probable as there is likely to be more interest in scenarios that are both probable and
desirable. This is not meant to imply, a priori, that there are necessarily fewer desirable than probable scenarios
(both sets are infinite from a modal point-of-view and given sufficiently fine gradations of differentiation and long
enough time-spans). The opposite might hold good, with the set of desirable being larger than the set of probable,
but as the point is to expand the set of choice relevant scenarios, it is, in any case, only the overlap that is of
interest in the end.
R. Miller / Futures 39 (2007) 341–362 351

Author's personal copy
By way of contrast, discovery of the unknown (knowledge that is neither tacit nor
explicit but must be discovered) and the application of the methods that combine creativity
and social science to the task of discovery usually do not involve the same methods or
competences as Level 1 FL. Certainly, both Level 1 and 2 FL entail learning processes and,
as such, share the principles of good pedagogy. However, in practice, the discursive
exposure of underlying expectations and values that works so well for developing Level 1
FL is in marked contrast to the much more reflective (often non-discursive) processes of
analytical refinement involved in Level 2 FL.
Without making any implausible claims to value and expectation neutrality, as our basic
interpretive structures are built upon values and expectations, the methods selected to
conduct Level 2 FL exercises must be at once amenable to extra-systemic/non-
extrapolative inquiry and wielded in ways that invite extra-systemic/non-extrapolative
imagination. In other words, the aim is to develop a model that has the most open set of
parameters or permutations for telling stories about the future (i.e., increasing the coverage
of Zone 5). The challenge is to develop a ‘‘space’’ for imagining possible futures. As with
any canvas or map, there are rules and conventions, rooted in specific ‘‘ways’’ of seeing the
world that circumscribe what is painted or projected into the frame. It is important to
make these explicit, as per point five (v) of the narrative rules outlined above. However, it
is equally important to justify, using theories and hypotheses, the selection of the variables
that describe the ‘‘possibility space.’’ Thinking of geography, it is obvious how even if the
territory on a map is unexplored it can still be located using latitude and longitude. In a
similar fashion a possibility space delimits, by setting out a coordinate zone, a range of
possible ‘‘locations’’ for stories of the future.
Picking up the thread of the late 19th century FS exercise about electricity used above to
exemplify Level 1 FL, here is a practical example of a four-step method for constructing a
‘‘possibility space’’ frame and then selecting specific points (scenarios) within that frame.
21
This example is meant to illustrate the concept without, in any way, exhausting the many
techniques for defining a possibility space. For the sake of brevity, there is no detailed
discussion of the premises and theory underlying the specific variables.
Step 1: First, select a subject,
22
in this case, continuing with a hypothetical exploration of
the future of electricity in the late 19th century, the subject that emerges from the Level 1
discussions is the pervasiveness of electricity (variable A)—with pervasiveness defined in
terms of how widely a technology is diffused (extent and diversity of use). When a
technology is invented and then commercialised, its degree of eventual diffusion remains
an open question. On the one hand, it is possible that it will not be picked up at all.
Alternatively, it might become widely diffused, entering all aspects of life—from the
workplace to the home.
Step 2: The second step is to develop a model, following the general precepts of social
science, that defines the attributes (variables) of the selected subject, in this case the
diffusion of electricity. Here, by way of example, the hypothesis is that the possibility of
diffusion (note this is not probability of diffusion) increases in line with two of the key
attributes of a technology: (a) how easy it is to use and (b) the number of different uses to
ARTICLE IN PRESS
21
Such modelling steps are common to social science methods and in the scenario technique literature as well [6].
22
In practice, Level 1 exercises can help with the selection and specification of the subject, even if the definitions
and metrics produced by Level 1 processes usually require subsequent ‘‘rigorous imagining’’ to push the
boundaries of the imaginable, using analytical refinement based on robust theories and models.
R. Miller / Futures 39 (2007) 341–362352

Author's personal copy
which it can be put [33].Fig. 2 describes the two variable spaces and the arrow illustrates
one hypothesised relationship (path) for a technology like electricity. As it becomes easier
to use and is applied to more different uses, the possibility of pervasiveness increases from
the lower left corner of the possibility space to the upper right.
Crucially, the probability of arriving at any particular point in this space remains
undetermined. No probabilities are attached to any particular point. As a result, possibility
spaces help to open up the field of points (outcome scenarios) and trajectories for getting to
the outcomes (path scenarios). It is one way of being systematic and explicit about the
hypothetical ‘‘what if’’. Forecasters also take this type of approach. Only their efforts,
including the theories and variables selected, focus on prediction and usually work within
the constraints of what is practically quantifiable.
23
The aim of a possibility space analysis
is to get beyond the predictive imperative by applying many of the same modelling
techniques to the challenge of expanding the set of possibilities for both building and
mapping scenarios.
This approach helps to overcome two drawbacks that often render scenarios less useful
for strategic purposes. First, this technique is careful to focus on Zone 1 (possibility) prior
to working on Zones 2 (probability) and 3 (desirability). This is critically important from a
strategic point-of-view because it provides a broader canvas (Zone 5) for imagining both
ARTICLE IN PRESS
Ease
of use
Simple
Difficult
Limited &
homogeneous
Unlimited &
heterogeneous
Range
of uses
Electricity
Fig. 2. Possibility space illustration—pervasiveness of electricity.
23
It is not the empirical testing that makes predictive forecasting models less appropriate to thinking about
possibilities. Rather, it is simply that the objective is usually to find a model that provides a good ‘‘fit’’ with past
data and on that basis offer probabilistic predictions about the future. Rarely is the aim to explore potential,
particularly ‘‘non-ergodic’’ change, as D. North points out [24]. By way of contrast, the goal here is not to forecast
(predict with any degree of probability) but to expand the range of potential scenarios (imaginable stories) based
on a more open framework for describing what may be possible. Hence the lack of quantification—at least
initially—is not only a virtue (it opens up potential) but in the case of transformative change a necessity, as the
categories of future quantification have not even been defined yet. Once the possibilities have been rigorously
explored, modelling based on quantitative and qualitative estimates of variables can be an important tool for
deepening the analysis of the factors that might influence rates and directions of change (for instance, see the
‘‘radar chart’’, p. 21 in [11], which is an example of metrics for a possibility space scenario of the learning intensive
society).
R. Miller / Futures 39 (2007) 341–362 353

Author's personal copy
the ends (goals for the future—Zone 3) and means (how to get there—Zone 2). This
facilitates efforts to articulate the emergent nature of the future. Second, a key tool for
pushing the imagination is the use of the analytical apparatus and findings of social
science. Not only do these hypotheses and models often (although not always) make it
easier to push the envelope of what is imaginable, even to the point that some possibility
space scenarios range outside Zone 1, but in addition, because the definitions and
frameworks abide by basic rules of social science, it is easier to connect an analysis of the
conditions for the realisation of particular scenarios to the assumptions about behavior,
institutions and power that shape current choices (action/policy).
Step 3: Having enlarged the set of possible futures for consideration when developing
scenarios, the third step is to select specific scenarios from what is still a vast space of
possibilities. The question is how? The criteria for selecting the scenarios will depend, in
large part, on the type of story (as per the narrative rules above) and the specific subject. Of
course, there are the Bear and GBU approaches that could be applied immediately to the
broader set of possibilities. These extrapolation and/or preference-based perspectives can
be used to make a selection from within the larger possibility space, either by taking the
starting point and rates of change as givens or by imposing a specific set of values for
differentiating end-points. However, once again, the use of values and predictive
parameters runs the risk of pre-empting imaginative and emergent options.
A more neutral and open-ended way of selecting scenarios from within the ‘‘possibility
space’’ frame is to focus on the functions and/or organisational attributes of the scenarios’
subject setting aside, for the moment and once again, the questions of probability and
desirability. This approach to selecting scenarios, which is just one among many,
24
has the
particular advantage of translating fairly directly into the functionalist/organisational
requirements of much decision-making, i.e., decisions that aim at achieving a certain
functional outcome (a specific product or end result such as an ‘‘educated’’ population),
using a particular form or organisation (like a factory or a school).
Continuing with the example of electricity, imagine that electricity is a technology that
has not yet traced its path across time (in formal terms a ‘‘counter-factual’’ scenario).
From this pre-electrical vantage point, imagine three hypothetical functions and two basic
organisational patterns that can be used to develop the counter-factual scenarios for
electricity: weapon (tool of war); local replacement for steam and water power in factories;
and autonomous source of power for all kinds of consumer products (most of which have
not been invented yet). Out of many possible ways of organising electrical power
generation, the two most contrasted options are: centralised and decentralised. These three
functions and two organisational options generate the six scenarios in Table 3.
Step 4: The next step is to take the six scenarios in Table 3 and project them back onto
the possibility space as per the mapping in Fig. 3. Obviously, the location of a particular
scenario within the possibility space is determined by how that scenario relates to variables
(a) and (b). Again, for the sake of illustration and without the in-depth analysis needed for
defining metrics and justifying the situation of the variables, scenarios S2, S4 and S6 are
ARTICLE IN PRESS
24
In addition to function/form criteria for selecting stories, there are other task determined parameters like
incentives/disincentives, signals/sensitivity, etc. What these criteria have in common is the aim of providing an
institutionally or organisationally non-specific continuum for the variables that define the coordinates of the
possibility space. This abstract, open-ended and non-probabalistic basis for selecting the scenarios is also a key
part of ‘‘rigorous imagining’’.
R. Miller / Futures 39 (2007) 341–362354

Author's personal copy
mapped higher on the grounds that decentralised generation implies that technical barriers
to use have been reduced—hence these scenarios score higher on ease-of-use. While
scenarios S4, S5 and S6 are deemed to exhibit a wider range of uses because, as a
decentralised tool for industry (S4) and a general tool for consumers (S5, S6), electricity is
assumed to be used in many different ways. In S1, where electricity is held exclusively by
the military as a specialised weapon dependent on the centralised generation of power,
there is little need to develop ease-of-use, while the range of uses is very narrow. Hence, S1
is in the lower left of the possibility space.
None of these scenarios can be considered strategic (informative for making choices) as
they only map function/form scenarios onto a technology pervasiveness possibility space.
Of course, we know that electricity ended up covering all three functions and, despite the
recent appearance of somewhat simpler and more efficient techniques for decentralised
power generation, the ease-of-use (and cost) problem was largely solved through
centralised provision of electric current. Getting to strategic scenarios, as illustrated in
Fig. 1, involves reconnecting with probabilities and preferences or Level 3 FL.
2.3. Level 3 LS: choice
Level 3 FL, building on the rigorously imagined scenarios of Level 2 FL, uses values and
expectations to assess today’s choices. Level 3 FL integrates the insights of the two
previous levels. From Level 1, it takes an awareness of values and expectations. From
ARTICLE IN PRESS
Ease
of use
Simple
Difficult
Limited &
homogeneous
Unlimited &
heterogeneous
Range of uses
Electricity
S2
S3
S6
S1
S4
S5
Fig. 3. Example of a function/form scenario for electricity.
Table 3
Form and function scenarios of electricity use scenarios
Function Organisation
Centralised generation Decentralised generation
Weapon Scenario 1 Scenario 2
Industrial Power Scenario 3 Scenario 4
Consumer Power Scenario 5 Scenario 6
R. Miller / Futures 39 (2007) 341–362 355

Author's personal copy
Level 2 FL it takes the discoveries of ‘‘rigorous imagining’’. The Level 1 values are crucial
for selecting between different scenarios in order to designate one (or an isovalue set) as the
strategic goal(s). Expectations are also a key ingredient for Level 3 FL because the
likelihood of making the decisions that matter for the strategic goal will depend, in large
part, on the expectational landscape. Level 2 imagination is crucial because it takes the
assessment of the present beyond the constraints of what is considered ‘‘realistic’’ (for
instance, the ‘‘I don’t dare hope for’’ or ‘‘it can’t be measured’’ that so often truncates
Level 1 thinking). Rigour is also critical because the analytically sound attributes of Level 2
scenarios means that the conditions for change are presented in decision relevant terms.
Level 3 FL builds on all of this capacity to think about the potential of the present and
provides the link to action.
Once again, an illustration may help. Continuing with the counter factual scenarios for
electricity, assume that the desired goal (the preferred scenario), as articulated for instance
by elected politicians, is to make electricity as pervasive as possible (a bit like information
technology today). This means that the policy challenge calls for an analysis of what it
would take to get to the upper-right of the possibility space (Fig. 2). Decision makers
would seek answers to a series of discipline spanning questions such as: What are the best
technical estimates that this is plausible? How practical are decentralised forms of electrical
generation such as, at that time, mini-water wheels, batteries, wind, solar? What are the
obstacles, in terms of people’s know-how, risk factors, cost, etc.? Can ease-of-use be
expected to improve dramatically without generating much higher costs (training, new
user-interfaces, etc.) How serious are the entrenched interests blocking certain decisions
and facilitating others?
Looking back, we now know from the history of electricity that some of the decisions
that locked in the form and function (the history) of electricity’s development were explicit,
but that many were not [33]. At a certain point, the path dependency or the cost of
changing course (technically, financially and from the point-of-view of overcoming
political obstacles) became very high [34]. Most of the decisive moves that led to the lock-in
of the centralised model were taken without any real awareness of the (what-if) roads not
taken. Historians now debate whether, at the time, there was either the will or the way to
explore the technological, economic and social possibilities—scenarios of plausible
alternative configurations [35]. Indeed were the terms (language), workplace tools and
methods and consumer goods and living patterns of ubiquitous electricity even
imaginable? Strikingly, much the same question can be posed today regarding the future
of energy or information technology or of the tertiary education sector.
To go into a little more practical detail, Level 3 FL involves four basic (and familiar)
steps:
(i) Strategic Goal: First the possibility space scenarios that are part of Level 2 FL need to
be looked at in normative terms. What is good or bad about these rigorously imagined
scenarios? Which one is preferred in a specific context? For instance there are many
different types of mixed economy and modern welfare state that across the developed
nations today—different ways of achieving similar outcomes that reflect the values
and history of each nation. Which one of these many different ways of achieving the
same degree of well-being is preferred even if, from a Pareto Optimal view, all of the
scenarios display equal levels of ‘‘well-being’’ or satisfaction (with the winners from
change compensating losers)? Here it is very useful to be able to reintroduce the values
ARTICLE IN PRESS
R. Miller / Futures 39 (2007) 341–362356

Author's personal copy
made explicit during the process of developing the Level 1 scenarios (note it is the
specification of values that is useful, not the Level 1 scenarios which have, by this
stage, been discarded).
(ii) Strategic Choices: The second step analyses the similarities and differences between the
selected story of the future and the story we tell about the present. The aim is to
identify the choices that are both ‘‘structurally’’ and ‘‘dynamically’’ critical (i.e.
without this particular change or continuity, the structure and/or the dynamics do not
hold together (work, function)).
25
How consistent or inconsistent are the choices that
underlie the strategic scenario with current decisions?
(iii) Strategic Probability: The third step, in light of the second, is to consider the
probability that the requisite decisions will be made. Given current expectations
(political positions, social conventions, attitudes, etc.), is it likely that the enabling
changes will occur? What do decision-making models, historical analyses and political
science tell about the play of power and conflict, inertia and volition when it comes to
the strategic choices?
(iv) Finally, the fourth step is to act on the basis of the strategic insights acquired in the
previous steps. Such an expectation may be too rationalistic, and in any case, will
depend greatly on the context. Still, the fundamental point of FL, and of the effort
required to identify scenarios that are at once possible, probable and desirable, is to
influence the choices people make. This last step is not at all self-evident. For futurists,
and also those who yearn for prediction, it is worth remembering the recent and tragic
example of predictions not heeded: the inundation of the city of New Orleans.
Predictive or non-predictive, a scenario is only a decision-making tool. Tools are not
always used, and neither are they always used for the ends the designers intended.
The role of FL and the HSS method for acquiring and practicing FL is an open one. By
way of a brief conclusion, I want to explore two opposing uses: planning versus
spontaneity.
3. Conclusion
The conception of education as a social process and function has no definite meaning
until we define the kind of society we have in mind.
John Dewey, Democracy and Education, 1915 [36].
Level 3 FL makes it easier to put current choices into strategic context in two ways.
First, more conventionally, it helps decision makers to question current goals in explicit,
actionable terms by drawing clear contrasts between the assumptions and content of the
present policy and the outcomes and preconditions of the strategic scenarios. Are the goals
the same? Are the functional relationships that make the scenarios plausible the same?
ARTICLE IN PRESS
25
At this stage, to keep the presentation simple, no distinction is made between comparative static (outcome)
and dynamic (path) scenarios. In practice, this distinction is crucial and thus the notion of structurally and
dynamically critical needs to be looked at in two ways—one for the structure of change and one for the structure
of systemic coherence. And then in terms of the relationship between the two, in particular in the context of what I
call ‘‘spontaneous society’’, where putting values into practice through spontaneous decision-making is what
determines the indeterminate future.
R. Miller / Futures 39 (2007) 341–362 357

Author's personal copy
What has changed to turn the potential into an operational, more practically realisable
vision? How dependent (contingent) are the changes depicted in the strategic scenarios on
changes in current policy and/or on new policy and/or on changes in underlying values,
individual and collective capacities, institutions, laws, cultural norms, etc. [44,45]?
Second, less contiguous with past practices, Level 3 FL may be one of the necessary
(although not sufficient) conditions for the emergence of a post-industrial society. In such a
society, if FL becomes quasi-universal like linguistic literacy today, then that society may
be capable of continuously exploring both the micro (individual) and macro (collective)
requirements for realising the potential of the present. This is more fundamental than it
might appear because building up this capacity may address one of the primary unresolved
problems with post-hierarchical governance (i.e. fuller democracy): how to reconcile
bottom-up local desires and power with top-down preferences and power? [50]. Or, framed
in terms of complex adaptive systems theory, if FL is an attribute of post-industrial
society, then it may help to improve the chances that the path taken by ‘‘blind’’
evolutionary processes is both consistent with people’s aspirations and makes the most of
the potential for innovation [37].
In this way, the strategic scenarios that are central to Level 3 FL offer a solution to two
critical ‘‘flaws’’ of non-hierarchical, complex self-organising systems: the non-teleological
nature of such evolution and the non-systematic discovery of potential for innovation or
alternative solutions. Society-wide acquisition of FL might provide a response to these two
deficiencies of evolutionary systems by creating the capacity, in the context of the type of
post-industrial society assumed at the outset of this article, to detect and engage the
potential of the present in four critical areas. First society-wide FL would help to equip a
society to embrace much greater dynamism (birth, death, entry, exit) and complexity
(multi-level, multi-purpose) of networking. In this context FL equips people and
communities (interest, practice) with an ability to invent and tell stories that can ascertain,
initiate and sustain the requisite common standards for network functioning. Second FL
develops an explicitness of values in practice that can turn spontaneous choices into the
realisation of our aspirations (without actually being able to know, project or want to
impose such aspirations on future generations). Third FL is a way of using the contingency
of outcomes (indeterminacy and volition) to shape/inform choices at the level of individual
and collective behaviour, rules, etc., in ways that embrace the ambiguity and complexity of
reality. And fourth, FL addresses the non-systematic aspect of blind evolution because it
offers a systematic examination of the opportunities (rigorously imagining the potential of
the present) and the need for innovation, where breakthroughs are waiting to happen and
where breakthroughs need to happen.
Level 3 FL shares many attributes of ‘‘strategic planning’’ without, however, becoming a
quest for predictive fulfillment. This may seem a small nuance and, if the opening
assumption regarding transition scale change turns out to be false, a largely irrelevant one.
But in order for FL to be effective in the context of a post-industrial ‘‘spontaneous
society’’, then it must remain resolutely non-predictive as well as strictly choice
contingent.
26
The non-predictive requirement stems from four basic attributes of a
ARTICLE IN PRESS
26
The principle of ‘‘choice contingency’’ is an assumption that does not deny that there are non-choice
contingent aspects of what happens, such as the sun rising, a meteor hitting the Earth or other ‘‘acts of the gods’’.
However, as accounting for such acts (including ‘‘wildcards’’) is either a contingency planning exercise or a blind
attempt to guess the unknowable, not a goal-seeking strategic reflection, it lies outside this discussion.
R. Miller / Futures 39 (2007) 341–362358

Author's personal copy
spontaneous learning intensive society: (i) heterogeneity, (ii) complexity, (iii) fluidity, and
(iv) spontaneity [13,31,42]. Under these conditions, prediction is a violation of principle.
There is neither a way nor add reference is there any need to know what will happen (what
people will want, do, have or think in the future).
27
What does matter is recognition of how
choice contingent the future is and, hence, why decisions made now needs to grasp, to the
fullest, the possibilities of the present (i.e. be futures literate).
Strictly speaking, there is no reason to violate the principle of ‘‘non-prediction’’ when
making choices meant to change what happens in the future. Everyone accepts the self-
evident proposition that the future is the outcome of a series of decisions that were always
made in the present, the here and now. The past time of human society stretches out
behind us in a long string of decisions (made or not made) and the future-time in a string of
decisions to be made. Following the logic of this ‘‘modal’’ assumption, there is no
necessary reason to impose predictive expectations on the choices made today as such
choices may or may not be followed by consistent or inconsistent choices tomorrow.
Furthermore, if a choice is made tomorrow that is inconsistent or in outright opposition to
the one made today, that is (according to one set of values) the prerogative of those
making the choices in the future.
A slight pause on this point is useful because it runs contrary to one of today’s ‘‘common
sense’’ principles enunciated by many and, perhaps most familiar, from Gro Harlem
Brundtland’s 1987 Report: Our Common Future, that defined sustainable development
‘‘as development that meets the needs of the present without compromising the ability of
future generations to meet their own needs.’’ Inherent in this principle is the belief that
today’s generations are in a position to determine what will or will not compromise the
ability of future generations to meet their own needs. In one interpretation, a common
one in the literature, this rests on a very strong predictive assumption regarding not
only the capacities but also the values of future generations. An assumption that is usually
left implicit.
Thinking about the relationship of the values past to values present, there are few who
would say that my choices today should be based on the assumptions made by my
grandparents about a women’s role in society, about the filial duties of the son, about the
way to ensure a livelihood, about the importance of the nation, about authority or
democracy. Yet my grandparents may have made choices, not only to ensure their survival,
but perhaps in their minds to perpetuate their way of seeing the world and living in it. As I
sit here writing, I cannot begrudge them either their conservatism (preservationism) or
their choices about how to survive in a fairly hostile context (early to mid-20th century).
What I can do is seek a way to continue their commitment to freedom and responsibility
by improving both my capacity and that of the society upon which I depend to put my
values into practice, into day-to-day life, into the string of choices that turn the present
into the future. It is also one of the profound criticisms of today’s world, one of the value
based reasons for striving to change it, that it is very difficult to reconcile many of our
ARTICLE IN PRESS
27
Claiming that it is possible to ‘‘know’’ what people will do or want (value) in the future runs contrary to both
‘‘history of the past’’ observations that the practices of daily life and values change over time, and the ‘‘history of
the future’’ assumption that it is probable that life and values will continue to change in the future. There is no
way to know what the content of those changes will be. Claiming that we want people in the future to have the
same values and do the same things as us is another story entirely. Unfortunately, these two different propositions
are often confused.
R. Miller / Futures 39 (2007) 341–362 359

Author's personal copy
aspirations with both the choices we face and the terrible reality of our times. Advocating
change is a moral imperative.
Which is precisely why FL, as a capacity, may be useful. FL embraces the profound
unknowability and contingency of the future by developing the capacity to imagine and
analyse the potential of the present in ways that are more consistent with our present
values. FL is neither about predicting what will happen nor planning the path from A to B,
but about improving the capacity to imagine and assess the potential of decisions made
now to create a future by putting values into practice. By combining an openness to the
potential of the future with a greater ability to invent stories that make sense of the present,
FL produces strategic insight without prejudicing the autonomy of people in the future to
see different options and hold different values. Still, FL is only a tool. And, like any tool, it
can be used for ‘‘good’’ or for ‘‘bad’’. Nor is the tool an end in itself.
Why, then, does FL matter? There is no way to know. What is clear, looking to the past,
is that the implications of generalised language literacy ushered in by universal compulsory
schooling, a hugely transformative initiative by any standard, made possible things that no
one could have imagined. Champions of schooling argued that formal education was the
only way to transmit and reproduce the workings of a complex society [36]. Is LS likely to
be as critical a tool for the 21st century as language literacy was for the 19th and 20th
centuries? Could futures literacy be a key enabler of a much more spontaneous, networked
and learning intensive society? Is this a way to reconcile the power of complex adaptive
systems with the aspirations of freedom and responsibility? The only way to find out is to
do it.
Acknowledgements
As with all such works, this paper has benefited greatly from the insights provided by
others. While exonerating them from any errors, I owe deep thanks to Madeleine Akrich,
Dan Atkins, Tom Bentley, Peter Bishop, Kristine Bruland, Hugues de Jouvenel, Laurent
Dominati, William Dutton, Pankaj Ghemawat, Michel Godet, Fabienne Goux-Baudi-
ment, David Hargreaves, David Hopkins, Esko Kilpi, Pierre Levy, Michael Mandel, Tony
Mackay, Nicola Meek, Wolfgang Michalski, Alain Michel, Isabelle Miller, Morris Miller,
Geoff Mulgan, Philip Van Notten, Jay Ogilvy, Erik Overland, Gilles Paquet, Helene von
Reibnitz, Jean-Claude Ruano-Borbolan, Ziauddin Sardar, Wendy Schultz, Hanne
Shapiro, Richard Slaughter, Mihai Spiriosu, Barrie Stevens, Michael Storper, Edna Yahil
as well as two anonymous referees. Lastly, I want to thank my students at the Masters in
Public Affairs at Sciences Po for their perception and perseverance.
References
[1] P.D. Aligica, Prediction, explanation and the epistemology of future studies, Futures 36 (10) (2003) 1027–1040.
[2] R.A. Slaughter, Where now for futures studies?, Futures 34 (3–4) (2002) 229–233.
[3] J. Dator, Advancing Futures: Future Studies in Higher Education, Praeger, Westport, CT, 2002.
[4] W. Bell, Foundations of Futures Studies: Human Science for a New Era. Vol. 1: History, Purposes, and
Knowledge, Transaction Publishers, Piscataway, NJ, 1997.
[5] M. Godet, Creating Futures: Scenario Planning as a Strategic Management Tool, Economica, PLACE, 2001
(M. Godet, Manuel de prospective strategique, Tome 1 and 2, Dunod, 2001, 2004).
[6] J.C. Glenn and the Futures International Group, Scenarios, in Futures Research Methodology, AC/UNU
Millennium Project, Version 2.0, /http://www.acunu.org/millennium/FRM-v2.htmlS.
ARTICLE IN PRESS
R. Miller / Futures 39 (2007) 341–362360
F

Author's personal copy
[7] C. Booth, M. Rowlinson, P. Clark, A. Delahaye, S. Procter, Scenarios and counterfactuals as modal
narratives, in: Paper to the Third Conference on Advancing Foresight Methodologies, Roskilde, Denmark,
12–14 May 2005.
[8] I. Prigogine, The End of Certainty: Time, Chaos and the New Laws of Nature, Free Press, 1997.
[9] K. Popper, The Poverty of Historicism, Routledge, 2002.
[10] R. Miller, The future of public sector reform: towards a learning society?, in: Adaptive Public Services:
The Next Phase of Transformation?, DEMOS & Hewlett Packard, London, 2003.
[11] R. Miller, T. Bentley, Unique Creation, National College for School Leadership, UK, 2003.
[12] OECD International Futures Programme’s 21st Century Transition series. (Available at: /http://www.oecd.
org/dataoecd/12/42/1903212.pdfS).
[13] R. Miller, Getting the questions right: challenges for 21st century policy makers, optimum online, Journal
of Public Sector Management 33:3 (Fall 2003). (Available at: /http://www.optimumonline.ca/article.
phtml?id=183S).
[14] K. van der Heijden, et al., The Sixth Sense: Accelerating Organisational Learning with Scenarios, Wiley,
New York, 2002.
[15] G. Ringland, Scenarios in Public Policy, Wiley, New York, 2002.
[16] J.A. Ogilvy, Creating Better Futures: Scenario Planning as a Tool for a Better Tomorrow, Oxford University
Press, Oxford, 2002.
[17] H. de Jouvenel, An Invitation to Foresight, Futuribles, 2004.
[18] P.W.F. van Notten, J. Rotmans, M.B.A. van Asselt, D.S. Rothman, An updated scenario typology, Futures
35 (5) (2003) 423–444.
[19] Cloudy Crystal Balls: An assessment of recent European and global scenario studies and models, European
Environment Agency, November 2000, /http://www.eea.eu.intS. P. Van Notten, Writing on the Wall:
Scenario Development in Times of Discontinuity, dissertation.com, 2005.
[20] Kennedy, A Guide to Econometrics, fifth ed, Blackwell Publishing, 2003.
[21] R. Slaughter, Knowledge Base of Futures Studies, vol. 2, 2005; W. Schultz, Defining Futures Fluency.
[22] D. Lewis, Counterfactuals, Blackwell, 1973.
[23] N. Rescher, Imgining Irreality: A Study of Unreal Possibilities, Carus Publishing, 2003.
[24] D.C. North, Dealing with the non-ergodic world: institutional economics, property rights and the global
environment, Duke Environmental Law and Policy Forum. X:1 (Fall 1999) 1–12. (Available at: /http://
www.law.duke.edu/shell/cite.pl?10+Duke+Envtl.+L.+&+Pol’y+F.+1S).
[25] K. Boulding, Basic Evolutionary Model, Evolutionary Economics, Sage, Beverley Hills, 1981, p. 44.
[26] National Education Association, The Future of Higher Education, /http://www2.nea.org/he/future/index.htmlS.
[27] National College for School Leadership, FutureSight, NCSL, 2004
[28] D. Istance, Schooling Scenarios, OECD, 2000.
[29] S. Vincent-Lancrin, University Scenarios, OECD, 2003.
[30] A. Wynberg, Wider and Deeper: A Review and Critique of Science and Technology Foresight Exercises
in the 1990s, AFI Monograph Series, 2003 /http://www.swin.edu.au/afi/publications/monographs/
index.htmS.
[31] R. Miller, Where Schools Might Fit in a Future Learning Society, IARTV, Victoria, Australia, 2003.
[32] R. Slaughter, The Knowledge Base of Future Studies, Professional Edition 2005, /http://www.
foresightinternational.com.au/books/kbfsproedn.phpS.
[33] K. Bruland, Technological Revolutions, innovation systems and convergence from a historical perspective,
University of Oslo, Convergence Project, February, 2001; C. Perez, Technological Revolutions and Financial
Capital, Edward Elgar, 2002.
[34] C. Perez, Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages,
Edward Elgar, 2003.
[35] D. Gugerli, The sociocultural aspects of technological change: the rise of the Swiss electricity supply
economy, Science in Context 8 (3) (1995).
[36] J. Dewey, Democracy and Education, Macmillan Company, 1916.
[37] E. Kilpi, Complexity and emergence, personal communication of the author; N. Johnson, The Science of
Diversity in a Fast Changing World, /http://www.lanl.gov/orgs/dvo/sci_diversity/overview.shtmlS.
[42] M. Buchanan, Complexity Tools in Evaluation of Policy Options for a Networked Knowledge Society, DG
Information Society, European Commission, June 2003.
[44] R. Miller, Measuring What People Know: Human Capital Accounting for the Knowledge Economy, OECD,
1996. (Available at: /http://www1.oecd.org/publications/e-book/9196031E.PDFS).
ARTICLE IN PRESS
R. Miller / Futures 39 (2007) 341–362 361

Author's personal copy
[45] D.E. Atkins, Revolutionizing Science and Engineering through Cyberinfrastructure: Report of the National
Science Foundation Advisory Panel on Cyberinfrastructure, National Science Foundation, February 3, 2003.
(Available at: /http://www.communitytechnology.org/nsf_ci_report/S).
[46] K. van der Heijden, Scenarios: The Art of the Strategic Conversation, Wiley, New York, 2005.
[47] The Millennium Project has produced a scenario bibliography with over 450 scenario sets available at
/http://www.acunu.org/millennium/information.htmlSwhich are organized in the ‘‘Futures Matrix’’ in the
following domains: Demographics and Human Resources; Environmental Change and Biodiversity;
Technological Capacity; Governance and Conflict; International Economics and Wealth; Integration or
Whole Futures.
[50] See Innovation UK /http://www.innovation.gov.uk/S.
[52] R. Slaughter, Futures Beyond Dystopia: Creating Social Foresight, Falmer Press, 2004.
[53] U. von Reibnitz, Scenario Techniques, McGraw-Hill, New York, 1988.
[54] P. Van Notten, Writing on the Wall: Scenario Development in Times of Discontinuity, Dissertation.com,
Florida, USA, 2005.
ARTICLE IN PRESS
R. Miller / Futures 39 (2007) 341–362362