Knowns and Unknowns in the War on Terror Uncertainty and the Political Construction of Danger Christopher Daase, Oliver Kessler

Introduction

HOEVER TURNED Donald Rumsfeld’s sentences from a press
conference into a poem, which circulated in the world press just
weeks before the start of the war against Iraq, had a fine sense for

the suggestive power with which the US Secretary of Defense encapsulated –
almost poetically – the central problem of foreign and security policy today:
uncertainty. According to Rumsfeld, there exist three forms of knowledge
and non-knowledge

that are relevant for political and military decision-

making. First, there is certain knowledge (‘known knowns’), on the basis of
which political programmes and military strategies can be developed.
Second, there is some knowledge about what we do not know (‘known
unknowns’), uncertainties that can be tamed analytically and reintegrated
into the decisionmaking process as calculable risks. Third, there is non-
knowledge about what we do not know and cannot know (‘unknown
unknowns’). This is the wild card, the unpredictable that can throw over the
most careful planning. Interestingly enough, however, Rumsfeld conceals a
fourth category of knowledge and non-knowledge: the knowledge we do not
want to know. These are the things we could know but rather decide not to
know by forgetting, suppressing or repressing them. Rumsfeld’s poem thus
needs another verse, which might run like this:

Finally, there are unknown knowns
The knowns
We don’t want to know.

The poem indicates that the widespread perception of political and military

decisions as being based on more or less firm knowledge is misleading.
Rather, knowledge and non-knowledge are equally constitutive for the
decisionmaking process. It is the relationship between what we know, what
we do not know, what we cannot know and what we do not like to know that
determines the cognitive frame for political practice. Scholars have long
acknowledged that, in international relations, uncertainty is endemic
(Schelling, [1960] 1980; Hughes, 1999; Cioffi-Revilla, 1998). But, after the end
of the Cold War, when old truths had lost their validity, the uncertainty was
acutely felt with regard to new international risks (Richardson, 1989; Beck,
1996; Daase, 2002). Since then, international terrorism has emerged as a
prime security concern that many claim requires bold new strategies because
of its shadowy character and its incalculable dangers.

This article investigates the knowns and unknowns in international security

412

Security Dialogue vol. 38, no. 4, December 2007

We use the word ‘non-knowledge’ as the antonym of knowledge in order to reserve the term ‘ignorance’

and its pejorative connotations for a specific form of non-knowledge.

Of course, we do not claim that this poem could not be extended differently. We thank an anonymous

referee for pointing out further alternatives and their normative implications.

and their impact on the ‘war on terrorism’. In short, we claim that, in order to
explain counter-terror strategies, it is crucial to understand how knowledge
and non-knowledge in international politics are conceptualized, assessed
and managed. The first part is thus conceptual. We show how different com-
binations of knowledge and non-knowledge determine the kinds of danger we
perceive. In short, we identify threats, risks, catastrophes and ignorance as
such kinds of danger. The second part of the article is methodological. We
demonstrate how different notions of probability are used to determine the
magnitude of danger we face. Here, we identify two different ways in which
non-knowledge might be given: structured and unstructured. While un-
certainty is treated as synonymous with risk in the former case, risk and
uncertainty are treated distinctively in the latter. As we will demonstrate,
these two notions of risk and uncertainty draw specific boundaries of politi-
cal responsibility. The third part of the article is political. We reveal how
concepts of danger and measures of probability are used to construct
counter-terror strategies and to justify action or inaction. In short, we discuss
deterrence, pre-emption, denial and negligence. We do not claim that these
strategies exhaust the list of measures taken, but we think that they are para-
digmatic for how knowledge and non-knowledge shape security policy.

Knowledge, Non-Knowledge and the

Conceptualization of Danger

There are many ways to categorize different forms of knowledge and non-
knowledge. Rumsfeld draws on a distinction that is widely used in the
sociology of knowledge and links two aspects: ontology and epistemology.
Both knowledge and non-knowledge have two sides. There is knowledge (or
non-knowledge) about things, and knowledge (or non-knowledge) about
ways to identify things. The first refers to the phenomena of reality as such,
which we might or might not know, thus having phenomenological or
factual knowledge or not. The second refers to the epistemological status of
such knowledge. We might or might not know methods of gaining know-
ledge. If we do know ways of escaping non-knowledge, we possess method-
ological knowledge.

Both kinds of knowledge and non-knowledge are independent, but linked.

We might have reliable methods of identifying observable facts, thus pro-
ducing known knowns. But, we might also have some methods of dealing with
phenomena that we are not 100% sure of, thus creating known unknowns. On
the other hand, we must accept that there might be things we do not even
dream of and have no method of anticipating; these are unknown unknowns.
But, there are also situations in which factual knowledge is available in

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

413

principle, but not used because it is ignored or repressed; this is the category
of the unknown knowns.

In the security field – traditionally defined

– there is empirical knowledge

about the ‘things’ that could pose a danger to the security of the international
system or to national security. Such ‘things’ could be political actors, such as
Al-Qaeda, of which we know (or do not know) that they exist and are prepar-
ing for terror attacks. Also, such ‘things’ could be military capabilities, such
as nuclear weapons in the Iraqi stockpile, of which we know (or do not know)
that they exist and can be used. Finally, such ‘things’ could be motivations,
such as Iran’s intent to acquire nuclear weapons, which we might or might
not know. Depending on how much we know about actors, capabilities and
motivations, our factual knowledge about security dangers is more or less
complete.

But, in the security field, there is also methodological knowledge about the

techniques for obtaining factual knowledge, dealing with uncertainty and
assessing danger. This knowledge consists of concepts, theories and analyti-
cal methods to interpret the – supposedly raw – factual data about the
dangers in the world. Conceptual models help, for example, to describe
collective actors, such as Al-Qaeda, despite their complexity as unitary
rational agents. More or less sophisticated methods are used to assess mili-
tary capabilities, such as in Iraq prior to the war, without having firm
evidence. Theories in the field of International Relations (IR) allow inferring
motivations, such as Iran’s, although direct information is not available.
Whether or not these methods are adequate is another matter. They are part
of the knowledge that is used to assess danger in international politics.

Combining empirical and methodological knowledge and non-knowledge

leads to the four forms of known knowns, known unknowns, unknown unknowns
and unknown knowns previously mentioned. Depending on the knowledge
we have, international dangers take on very specific forms. If the empirical
facts of a danger are known (i.e. if an enemy actor exists with malign inten-
tions and a sufficient military capability to do harm) and if this knowledge is
known to be reliable, a threat exists. This was paradigmatically the case
between the USA and the Soviet Union during the Cold War. If, however, the
factual knowledge is partial, yet methods exist for reducing the uncertainty,
we might speak of security risks. Such risks have been perceived after the end
of the Cold War with regard to nuclear proliferation, terrorism, organized
crime and similar issues. If no (or scant) factual knowledge about a danger
exists and a method for assessment is not available, the danger, if material-
ized, will appear as disaster. If, finally, the facts of danger are largely known
but, for whatever reason, this knowledge is neglected, suppressed or forgot-
ten, the danger is intensified by what we call ignorance.

414

Security Dialogue vol. 38, no. 4, December 2007

We concentrate here on international and national security, although we acknowledge that societal and

human security are also important.

Probability and the Assessment of Danger

As much as those four forms of danger are based on particular methodologi-
cal and empirical knowledge – or non-knowledge – they are rather silent
on how they are identified, assessed and managed. However, for security
policy, the assessment of dangers is essential. There are many dangers in
international politics, but actors – be they national governments or interna-
tional organizations – have to set priorities and act on estimates of which
danger could do the most harm for the most people. Like all politics, these
prioritizations are contested, but, nevertheless, they build the basis for
political and military decisionmaking. How, then, are dangers assessed in
international relations? During the Cold War, threat assessment was relative-
ly easy. The enemy was known, its military capability was identified and its
intentions understood – or so it was believed (Baldwin, 1971; Knorr, 1976;
Cohen, 1979). This confidence has, however, disappeared. In today’s world
of asymmetric warfare and the breakdown of stable interpretation patterns,
many more unknowns enter the equation – with the consequence that the
type of knowledge and non-knowledge relevant for formulating security
policy changes. In this overall transformation of security policy over the last
decade, the concept of probability has gained prominence in security studies
in assessing the magnitude of danger.

Although the notion of probability

certainly conveys an element of contingency, the literature dealing with it,
however, assumes that probabilities themselves refer to objective and meas-
urable sequences or regularities. Probabilities are conceptualized as a specific
kind of knowledge allowing for secured knowledge of known knowns.
Consequently, this literature ultimately disregards the way in which the
unknown equally structures practices. As probabilities conceptualize the
very boundary of knowledge and non-knowledge and the way in which
the latter is translated into knowledge, this turn to a probabilistic vocabulary

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

415

Empirical knowledge

‘knowns’

‘unknowns’

‘known’

Methodological knowledge

‘unknown’

Figure 1: Four kinds of dangers

threat

risk

ignorance

disaster

See, for example, Wills (2005).

in assessing the magnitude of dangers nevertheless provides some indication
of the manner in which the mix of knowledge and non-knowledge changes.

To unveil the constitutive role of non-knowledge for probability theories,

let us look for a moment to the 17th century, where Scholasticism still deter-
mines basic convictions about the physical world, science and the relation of
human existence to God. In this context, the Platonic distinction of idea and
phenomena, of essence and existence, structures the scientific vocabulary
and divides scientia from opinio. Scientia, classical metaphysics, is the search
for the being of being, for the discovery of universal and necessary laws. This
quest for God, for the first principle in which all necessity is located, deter-
mines the boundaries of legitimate inquiries and methods. For the use of
proofs and justification for claims and propositions, therefore, it is necessary
to also take on the form of ‘demonstration’, that is, the application of right
categories and reference to ‘old’ authorities.

In contrast to this world of

primary qualities and direct knowledge, the lower sciences had to analyse
secondary qualities, signs and symptoms to recognize possible causes.
However, in this world of ‘phenomena’, there can only be opinio. There can
only be the approval of an argument, but not demonstration. Within the
lower sciences, probabilis characterizes this approval exactly. It points to a
conviction, aestimatio, resulting from deliberation, argumentation and dispu-
tation (Hacking, 1975: Chapter 2). Still, for an opinion to be ‘probable’, it
needs to be supported by authorities.

However, the social conditions under which this pre-modern understand-

ing is based changes dramatically during the 17th century. Characteristic of
the Scholastic world was a world-view based on the distinction of aeternitus
and tempus, where the stare and eternal realm of God and ‘being of being’ is
juxtaposed against human existence ‘in time’. Within these confines, the
scientific vocabulary of proof, evidence, demonstration, etc. is seen to be
characterized by a predominance of the past and the quest for the ‘original’
(see also Hacking, 1983). In the 17th century, this distinction of aeternitus and
tempus is increasingly replaced by the triad past, present and future.

Via this

change, the past loses its capacity to determine the scientific vocabulary as
the idea of science, because the revelation of the ultimate truth is abandoned.
Increasingly, science is understood as a ‘discovery’ process of predominantly
technical knowledge and is thus linked to the progress of society in general.
In this overall transformation of the temporal conditions of human life, a

416

Security Dialogue vol. 38, no. 4, December 2007

Within the Scholastic tradition, knowledge was grounded in authorities. Books were not seen as simple

texts, but as ‘speaking’ persons. For an analysis, see Luhmann (1997).

There is a dominance of the past over the future, which can be seen in the justification of dynasties; see

Koselleck (1979).

See Esposito (2002). Of course, the future existed before the 17th century, but the distinction of past/future

became the dominant form of observing the temporal condition of human existence. See also Luhmann
(1991: 23).

redefined concept of nature ultimately opens the door for probability.

There

is no need to consult the old authorities eternalized in their books; nature can
now be asked directly. The probabilis of signs turns into causes and becomes
the possibility to read nature’s laws.

The emergence of probability, in other words, parallels an overall trans-

formation of societies manifested in the discovery of the subject, social con-
tract theory, the invention of positive law and subjective rights, and thus a
particular modern notion of knowledge and understanding of how the world
might be known. Ultimately, explanations of possible harm by sin or Fortuna
are replaced by contingency and thus uncertainty and risk.

As Mary Douglas

(1992: 26) once said, ‘risk provides secular terms for rewriting scripture: not
the sins of the fathers but the risks unleashed by the fathers are visited on the
heads of their children’.

12 13

What is of less interest at the moment is the institutional impact of risk in

the form of insurances,

or the way in which the exact meaning of risk and its

relation to rationality, responsibility and blame changed over the centuries.

Rather, what is of interest at this point is the particular way in which the
semantic field of probability emerged as it gave rise to a specific ambiguity
concerning the category of non-knowledge. Let us go back to Leibniz for a
moment.

Leibniz shows that the notion of equal probability can have two

meanings. His first description is based on the principle of sufficient reason,
where he uses notions of opinio and aestimatio, two categories of Scholastic
epistemology. The expression that two cases have equal probability expli-
cates a degree of belief due to insufficient knowledge. In his second argu-
ment, Leibniz uses probability as degree of possibility, as aeque faciles seu
aeque possibiles. Leibniz uses facile, which refers back to Aristotle’s notion of

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

417

For introductions to probability theory, see Cohen (1989); Good (1987, 1950); Benenson (1984); Weatherford

(1982).

It would be interesting to analyse this fusion of several semantics during the 17th century more deeply: the

ascendancy of science, the emergence of probability, and risk. At the same time, risk is linked to a rede-
fined understanding of necessity and security. We can only speculate that the Thirty Years War functions
as a catalyst where the meaning of the past, the human and eternal life change, accompanied by the
demise of the church as a unifying force, in which knowledge increasingly moves from monasteries to
universities. From this perspective, risk is irremediably linked to the rise of the territorial nation-state. For
a discussion, see Kessler (2007).

The roots of risk go back to antiquity, and, although the term did not exist at that time, the functional

equivalent of today’s risk is antiquity’s explanations for possible harm in terms of myths, gods and ora-
cles. For a historical analysis of risk, see Luhmann (1991: Chapter 1); Bonss (1995); Ewald (1991); Bernstein
(1998).

Rooted in probability theory, the basic metaphor becomes the bet, or gambling, where references to sin, the

will of God or other rather esoteric explanations become illegitimate.

As Robert Castel (1991: 289) explains: ‘a grandiose technocratic rationalizing dream of absolute control of

the accidental, understood as the irruption of the unpredictable . . . a vast hygienist utopia plays on the
alternate register of fear of security, inducing a delirium of rationality, an absolute calculative reason and
a no less absolute prerogative of its agents, planners, and technocrats, administrators of happiness for a
life to which nothing happens’.

Ewald (1991: 199).

Douglas & Wildavsky (1982).

Leibniz, quoted in Hacking (1975: 125).

potentia. Equal probability does not refer to a lack of knowledge, but to future
states of the world. This understanding is further developed by Laplace and
his definition of probability as relative frequency.

These two very distinct and very modern definitions of probability – that is,

as relative frequency or degree of belief – are based on two definitive notions of
non-knowledge and uncertainty. The definition of probability as a relative
frequency characterizes uncertainty as a probability distribution over possi-
ble states of the world. For this setting to apply, uncertainty needs to be
characterized in terms of a distinct and exhaustive set of possible states of the
world and a probability distribution over this set. A die can only show a
number from one to six – and the ‘category’ of ‘three’ is different from ‘four’.
This framework is commonly known from expected utility theory and
mathematical risk analysis, where risk is embedded in broader assumptions
of rationality and axioms of probability calculus. As this literature shows,
and exactly because contingency is understood to be an ontological concept,
probabilities mirror objectively given facts. They provide a kind of secured
knowledge about possible states of the world.

A second meaning of uncertainty emerges within the epistemological defi-

nition of probability as a degree of belief. Consider a moment, a situation
where a person is located in a lit room – and then somebody switches off the
lights. In this dark and unfamiliar room, one has to find out what objects are
actually in the room and where they are located. The probability of what and
where an object is is not subject to a frequency, but rather to a degree of belief.
Probability does not constitute knowledge of the world; rather, it constitutes
knowledge about our engagement with the world and includes qualitative
judgements about ‘what is the case’. In this setting, uncertainty – that is, the
kind of relevant non-knowledge – is at first unstructured and needs to be
absorbed via evidences or processes of trial and error into manageable risk.

In this context, where uncertainty is juxtaposed to risk (Knight, 1921;

Keynes, 1937: 213–214), uncertainty changes the condition of possibility – or
impossibility – of action. It is subject to a different rationality, based on norms
and values and not on instrumental criteria. These conventions and norms
thus represent a different kind of knowledge, and at the same time constitute
the limits of a potential applicability of formal reasoning and the insurability
of the future. From the particular way in which uncertainty is translated into
manageable risks, one might think of hierarchies or institutions as other alter-
natives; however, this is not of particular relevance at the moment. More
important are the methodological implications these two notions of uncer-
tainty and risk have. Probability theories provide different conceptualiza-
tions of how the future and the present are connected via characterizing the
unknown. Risk names the boundary of what an individual can and does (not)
know, what lies in his responsibility or what is subject to, for example, a
‘higher force’.

418

Security Dialogue vol. 38, no. 4, December 2007

Politics and the Logics of Danger

The primary focus on knowledge and rationality within security studies
neglects the constitutive function of non-knowledge for dealing with the type
and perceived magnitude of danger. Having outlined two very distinct
meanings of uncertainty, we can now analyse how various kinds of non-
knowledge manifest themselves in specific security measures. A definition of
uncertainty in terms of risk already presupposes a well-defined ontology
within which knowledge and non-knowledge are assumed to be mutually
excluding. The crucial point, however, is not that an empirical fact is
assumed to be either known or unknown, but that dangers are always con-
ceptualized in terms of a ‘lack’ of knowledge. Dangers manifest themselves
owing to incomplete knowledge.

As soon as uncertainty is treated distinctively from risk, then non-know-

ledge and knowledge are not mutually exclusive, but mutually implicating:
we know because we do not know. The glass is not either half full or half
empty. Rather, as soon as we gather more information and increase our
‘stock’ of knowledge, our non-knowledge also increases and we become
more aware of what we do not know. Such a position recognizes a funda-
mental uncertainty, as there is always contingency in the particular way in
which uncertainty is absorbed into manageable risk. There are always other
perspectives, other worlds and other models that could frame the issue-
area differently. These two principled conceptualizations of non-knowledge
allow us to frame the current fight against terrorism as a clash of the different
logics of dangers, including different kinds of non-knowledge and security
policies. In the following, we will discuss the known knowns, known unknowns,
unknown unknowns and unknown knowns in terms of deterrence, risk manage-
ment, catastrophe and ignorance, respectively.

Known Knowns

The known knowns of security politics represent the realm of available secured
knowledge in political decision processes. This is the paradigmatic case of
structured uncertainty that we encounter in relative frequency probability
theory and mathematical approaches to security studies. This view assumes
a well-defined ontology, that is, the existence of recursive patterns of ‘law’-
like regularities where our knowledge mirrors empirical facts.

This knowledge consists of the use of a particular vocabulary within a given

framework, that is, the concepts and special notions that demarcate the frame
for evaluating international political situations. Of course, how one names a
situation has crucial consequences for the political processes that follow. It
matters whether a group is named as ‘terrorists’ or ‘freedom fighters’, ‘guer-

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

419

rilla warriors’, ‘combatants’ or ‘illegal combatants’. And, of course, these
concepts also transport evaluations and judgements – for example, the ille-
gality of ‘aggressions’, the despicableness of terrorism and the desire for
‘development’. However, mathematical approaches to risk management can
only work when these questions have been settled beforehand.

This ‘secure’ knowledge becomes more complex when it is conveyed by

analogies. Particularly, historical analogies are of central importance for
political processes in evaluating situations (Khong, 1992). Words like
‘Munich’, ‘D-Day’ or ‘Vietnam’ are not simple descriptions of historical facts
in the past, but convey particular lessons of past experience for today’s prob-
lems. ‘Munich’ became the analogy for a failed appeasement policy that only
increased the appetite of the maniac. It transports the insight that appease-
ment does not work on dictators and may not avoid but actually worsen the
situation. Analogies represent the biggest part of the known knowns in crisis
situations. If one looks at the use of historical analogies in journals like
Foreign Affairs (Jones, 2004), one finds that their use and the gravity of the
associated crises are positively correlated, because the analogies represent
codes for secured knowledge.

The realm of secured knowledge becomes even more complex, and is

already heavily contested in its universal claims, when (bounded) causal
if–then relations are included (Smoke & George, 1973; George, 1993). Such
knowledge is often constructed with mechanical pictures in mind. Consider,
for example, the picture of a balance of power in which necessarily competi-
tive power-centres develop and give rise to hegemonic wars. Or the view,
suggested by the scapegoat theory, that war is often started because external
confrontation helps to cut off and paper over domestic conflicts. Alter-
natively, think about the idea that only the right combination of sticks and
carrots will guarantee the success of diplomatic negotiations.

Finally, this category also contains the knowledge of facts that we find

located in encyclopaedias, news agencies, research institutes, international
organizations, embassies, civil and secret services, etc. This is a kind of
knowledge that can be collected and ‘brought’ together, and, even though it
is not always readily at hand, it is nonetheless in principle always available.
On the other hand, as the reference to probability theory shows, the founda-
tions of the known knowns crack open when basic concepts of political dis-
courses are contested, when analogies do not work and mechanical if–then
sentences are invalidated – in other words, when epistemological or diag-
nostic considerations need to be reconsidered. Transformative crises, such as
the end of the Cold War or 9/11, are such cases, where available and secured
knowledge becomes inadequate.

420

Security Dialogue vol. 38, no. 4, December 2007

Example: Deterrence

Deterrence can serve here as a paradigmatic example for the particular kind
of (non-)knowledge of known knowns for three reasons: First, it was enforced
only at the onset of the Cold War (Freedman, 1989); second, it represents a
particular way of dealing with political knowledge deemed insufficient (e.g.
in relation to the motives and the military capacity of the USSR) (Schelling,
[1960] 1980); and, third, it became one of the known knowns of foreign and
security policy during the Cold War (Zagare & Kilgour, 2000). After the end
of the East–West conflicts, and in particular with the start of the ‘war against
terror’, the concept of deterrence has lost much of its usefulness, although
politicians and commentators like Rumsfeld and Wolfgang Schäuble have
repeatedly made demands for a ‘new form of deterrence’.

The basic mechanism of deterrence – in both its nuclear and its convention-

al form – is to avert an aggressive act by an opponent by threatening that
opponent with unacceptable losses should it carry out the act (George &
Smoke, 1974). For example, one could deny all the potential benefits of a
potential aggressive act (deterrence by denial) and/or one could threaten to
heavily punish unacceptable behaviour (deterrence by punishment) (Snyder,
1961). For this strategy to work, both the deterring and the deterred state
need to know what the potential military, political and propagandistic
benefits of their behaviour are, and what particular punishment might be
highly sensitive. That is, the cost and loss functions of political actions, the
epistemological question of qualities, of context, and the determination of the
‘currency’, need to be established in advance.

This presupposition of fixed categories is also valid – cum grano salis – for

considerations on how to deter terrorists (Davis & Jenkins, 2002). However,
and in contrast to the framework of Mutually Assured Destruction (MAD),
the empirical and methodological knowns are not easily available. One of the
most frustrating elements of the terrorist attacks on New York and
Washington was precisely the non-existence of an identifiable actor and the
lack of any clear political end that could be served with the attack. It is
extremely difficult to identify or deny a potential gain if there is neither an
actor nor a common currency (Münkler, 2002) against which threats and
claims can be evaluated. Whereas somebody might deny the exchange of
prisoners to reject a potential benefit, the symbolic power of the tumbling
World Trade Towers cannot be ‘denied’. Consequently, an analogy between
MAD and terrorism is highly problematic. One cannot deter terrorists in the
way that one might deter states – particularly if the ultimate threat, death, is
not perceived as a threat but rather welcomed by suicide terrorists. The
benefit of the act is thus not some ultimate goal, but the act itself.

Regardless of the aforementioned problem, the network-based organiza-

tion of terrorist groups escapes the logic of deterrence (White, 1998: 32–43).

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

421

Terrorists have no address to which a note could be sent. As a network, they
may have a leader but no government, no territory, no ambassador and no
people. Terrorist networks are barely institutionalized, and their differentia-
tion and organizational logic work horizontally rather than vertically. Even if
parts of the network might be destroyed, the network itself can regenerate
itself rather quickly. In this context, the logic of deterrence by military retali-
ation and physical harm might prove to be extremely difficult to enforce
(Gotowicki, 1997). Moreover, as the network is organized globally, how can a
state retaliate against a network that uses the state’s own territory as a base
for its operations?

This might be the reason why states are often inclined to orient themselves

with actors who are like themselves, in order to focus on states that ‘harbour’
terrorists, so-called state sponsors of terrorism. The assumption that states
are responsible for prohibiting terrorist activity directs the focus. It is
believed that, as long as states and not the network itself are addressed, tradi-
tional diplomacy works. States have an address, at least. The difficulties in
deterring terrorists and the decision to focus on ‘sponsor states’ show that
deterrence as a specific security policy is based on secured knowledge about
the perceived threat, on methodological and empirical knowns (known
knowns). Most of all, for deterrence to work, there needs to be a threat, a
reason why the threat emerged and a currency for evaluating the threat. But,
how can one deter terrorists who want to die and whose rationale for death
and destruction is not of ‘this’ world?

This discussion can be summarized as follows: threats are dangers in the

status of known knowns insofar as the parameters of their evaluation need to
be common knowledge. A threat for actor A exists if three conditions apply:
first, there is another actor, B, that can be identified as such; second, an inten-
tion of actor B needs to be recognizable and pose a risk of harm to actor A;
third, there is a potential instrument available by which actor B can inflict
some considerable damage on actor A (Daase, 2002: 15; Singer, 1958;
Baldwin, 1971; Knorr, 1976; Cohen, 1979; Walt, 1987). This does not mean that
there might not be some misperceptions at play (Jervis, 1985; Stein, 1989),
emerging from psychological distortions and false information. Instead, it
means that, in principle, the level of analysis and the processes by with the
threat is analysed need to be known.

This is why, during the Cold War, the security discourse took place within

relatively stable parameters: there was almost no doubt about the existence,
the intentions and the military capacity of the opponent. Controversies
existed, of course, but these were mainly on details. For example, there was
debate over the level of unity of the Eastern bloc, the exact nature of its inten-
tions (coexistence or expansion) and its exact potential (Is there a missile
gap?). In principle, even on these points, there was some kind of agreement
on how these questions – theoretically – could be answered and the missing

422

Security Dialogue vol. 38, no. 4, December 2007

knowledge generated. This overall epistemological and ontological frame-
work was the basis for the security consensus during the Cold War. A
security discourse on threats is possible as long as there is a relatively com-
mon agreement on the description of the situation, and the chances for a
coherent security strategy are given (Pruitt, 1965).

Known Unknowns, or the Art of ‘Traditional’ Risk Analysis

The threat paradigm of security politics might still be valid today in some
parts of the globe. There are still some states who clash over territory and
access to resources, and that pose military threats to each other. But, the
kinds of dangers the global community faces are increasingly of a different
kind. They often lack an identifiable actor, a hostile intention or the promi-
nence of military capability. That is, we are now dealing with unknowns. Of
course, politics is shot through with ‘unknowns’, and the concept of uncer-
tainty is not a new invention within security studies. In a sense, uncertainty
has characterized political thought since Machiavelli and Hobbes, either in
the form of fortuna or in the description of the state of nature. That uncer-
tainty is thus an essential, and not only an accidental, problem of politics
might not be heavily disputed. However, uncertainty as used in political
science is defined in terms of risk. As we discussed above, it is understood to
be an ontological concept that signifies a realm between necessity and the
impossible. ‘God plays dice with politics, but the results are ordered, not
random’ (Cioffi-Revilla, 1998: 11). In this context, the social theory of risk
tries to develop means and methods of dealing with the unknown – to provide
instruments and procedures from which the unknown can be made known,
where, in the end, uncertainty can be incorporated as risk in decision pro-
cesses. Risks, in this narrow sense, are thus emerging threats with the status
of known unknowns, that is, calculable uncertainty (Knight, 1921).

Calculating risks does not mean that they can be measured objectively.

Not all uncertainties are of a quantitative nature, and thus understandable
within the common definition of rationality (Bonss, 1991; Kadvany, 1996). In
particular, the identification and assessment of risks may vary according to
political interests or cultural context (Douglas & Wildavsky, 1982). In these
confines, the traditional concept of deterministic causality has lost its valid-
ity. Uncertain political results and uncertain strategies do not follow social
laws, but – at most – probabilistic laws. The best that political research can
strive for is thus probabilistic knowledge, knowledge regarding necessary
and sufficient reasons and causes, which will surely not be able to predict
single events, but to point to trends and tendencies through the identification
of conditions under which their realization is likely.

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

423

Example: Terrorism

In this context, the question of how big the threat of international terrorism
currently is can be conveniently answered by pointing to the next possible
act. The rationale behind many counter-terrorist acts is the constant threat of
a subsequent act – which is indubitably anticipated for some point in the
future. Within such a point of view, the current calm is only the calm before
the next storm. Taken seriously, this statement is both true and trivial.
However, it is exactly the triviality of this insight – that is, that a next terror-
ist ‘attack’ will happen – that allows proponents to structure and shape the
contemporary security policy discourse. There are a couple of reasons for
this. Most basically, there are two standard models for examining the risk of
terrorism – which are, however, both inadequate (Falkenrath, 2001). First,
there is the inquiry into the motivational structure and the extrapolation from
past terrorist activities. Second, there is the attempt to calculate the possible
risks from the expected losses and the probability of a certain state of the
world occurring. The former is preferred by terrorist and regional experts,
the latter by security practitioners and security experts.

The problem of the first method is that it cannot trace new developments

and spontaneous changes in the motivational structure. There is always a
first time, one could argue. Even the act of hijacking planes in order to
destroy skyscrapers was not ‘predictable’. Such behaviour was simply not on
the screen; it was unimaginable. These methods of explorations are thus
inherently conservative, and they systematically underestimate the associ-
ated risks. The problem with the second method is that nobody can actually
‘calculate’ the loss of thousands of lives.

If the risk of terrorism is defined in

common terms of probability and potential loss, then the focus on terrorist
‘world events’ consequently leads to a reduced importance of probabilities.
The improbability of the risk’s manifestation becomes irrelevant, as the costs
would reach infinity (Jenkins, 1999). The classic calculation of risks from
terror thus tends to overestimate and to dramatize terror. This overestima-
tion of risks is not only directed at the level of risk assessment but includes
the formulation of a prudent ‘risk policy’: if one factor of the risk equation
goes to infinity, a fair representation being a terrorist attack with nuclear
weapons, then there is no rational measure for anti-terrorist measures.

This point is close to a further characteristic of risks: their political nature.

For example, in 1998, the Rumsfeld Commission criticized the Clinton
administration for systematically playing down the security threat posed by

424

Security Dialogue vol. 38, no. 4, December 2007

Though there certainly are attempts to value human lives. And, even though policymakers might use

these attempts, there are two objections. First, it is almost trivial to see that the value of human lives
changes from policy issue to policy issue and country to country. Second, and in spite of these attempts,
it seems to be a far cry to a situation where the ‘loss’ of an attack will be presented as a ‘bill’ due next
month.

missiles, and the intelligence services for working with the wrong method-
ologies (Rumsfeld Commission, 1998: Sections II.A, II.G). In a sense, this was
basically a battle of methods within the US intelligence community, about the
adequate methods for political risk analysis: which criteria allow for pruden-
tial recognition and evaluation of an emerging threat, allowing for effective
counter measures? Back in 1995, the National Intelligence Estimate had
declared that (as summarized in a statement by the chairman of the National
Intelligence Council, Richard Cooper):

the Intelligence Community judges that in the next 15 years no country other than the
major declared nuclear powers will develop a ballistic missile that could threaten the
contiguous 48 states or Canada. (Cooper, 1996)

The Rumsfeld Commission (1998: Section II.A) report, on the other hand,
emphasized that:

concerted efforts by a number of overtly or potentially hostile nations to acquire ballis-
tic missiles with biological or nuclear payloads pose a growing threat to the United
States, its deployed forces and its friends and allies. . . . [North Korea, Iran and Iraq]
would be able to inflict major destruction on the U.S. within about five years of a
decision to acquire such a capability (ten years in the case of Iraq).

Three years after the 1995 National Intelligence Estimate, the Rumsfeld
Commission had thus drastically reduced the time horizon of a potential
threat and increased the level of perceived risk. The main idea of the report,
however, was to criticize traditional methodologies of threat analysis in the
defence community:

In analyzing the ballistic missile threat, the Commission used an expanded method-
ology. We used it as a complement to the traditional analysis in which a country’s
known program status is used to establish estimates of its current missile capabilities.
We believe this expanded approach provides insights into emerging threats that the pre-
vailing approaches used by the Intelligence Community may not bring to the surface.
(Rumsfeld Commission, 1998: Section II.G.)

Although the Rumsfeld Commission did not provide any probability assess-
ment of the imagined developments, its results were taken by missile defence
advocates as evidence that the US intelligence community would misjudge
the situation and severely underestimate the danger. The scolded National
Intelligence Council subsequently embraced this new method to emphasize
the possible and not the probable. The subsequent Foreign Missile
Developments and the Ballistic Missile Threat to the United States Through 2015
thus reads:

North Korea could convert its Taepodong-1 space launch vehicle (SLV) into an ICBM
that could deliver a light payload (sufficient for a biological or chemical weapon) to the
United States, albeit with inaccuracies that would make hitting large urban targets
improbable. North Korea is more likely to weaponize the larger Taepodong-2 as an ICBM

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

425

that could deliver a several-hundred kilogram payload (sufficient for early generation
nuclear weapons) to the United States. Most analysts believe it could be tested at any
time, probably initially as an SLV, unless it is delayed for political reasons. (NIC, 1999)

We quote this passage to show how an emphasis on the possible rather than

the probable, and the almost complete lack of political risk assessments in the
1999 National Intelligence Estimate softened the established criteria used by
intelligence services to evaluate threats (Cirincione, 2000). The ‘growing
threat’ through the proliferation of carriers was based less on a dramatic
change of the armament behaviour of Iran, Iraq or North Korea, and more
on the change of the political parameters of US threat assessment. Put more
precisely, it was a politicization of risk analysis.

One can see the traces of this battle of methods even in the controversies

over how to explain the misinterpretation and false estimation of the intelli-
gence associated with decisionmaking concerning the Iraq War. As US
weapons inspector David Kay (2004) noted after the long and ultimately
unsuccessful search for evidence of a renewed weapons programme in
Saddam Hussein’s Iraq: ‘We were almost all wrong.’ In this context, it is
interesting to note that, a few weeks before the new war, US Vice-President
Dick Cheney (2003) argued that giving the inspectors more time would be
wrong, as more procrastination would increase the military risk posed by the
Iraqi regime: ‘The risks of inaction are far greater than the risk of action.’ This
thought served as central rationale for the military preventive strike: military
action was essential now to prevent future potential loss. This reasoning was
also behind the ‘pre-emptive strike’ doctrine by which the US government
undermined Article 51 of the United Nations Charter and the ‘Caroline cri-
teria’. Without mentioning the conditions under which the supposed risks
are supposed to grow, Cheney’s assertion is as empty as the sentence that
the next terrorist attack will surely come. This sentence nicely illustrates the
argumentation of the Bush administration that the cognitive uncertainty of
the known unknowns was not seen as a problem to be solved, but as the justi-
fication for military action. Cheney’s argument is more than just a simple
assertion – it is a programme for how the USA will deal with non-knowledge
in international politics.

Unknown Unknowns and the Logic of Catastrophe

Rumsfeld’s reference to the unknown unknowns was used as exculpation, an
apology for the fact that the decisions of the Bush administration were made
in a situation where necessary information was not available and ‘security’ of
its judgements unattainable. Here, and in contrast to the ‘risk as uncertainty’
definition we encountered in the realm of known unknowns, there is no meas-
ure to tame uncertainty

In this sense, uncertainty is not a part of the decision

problem in the form of risk management, but it is –and inevitably always will

426

Security Dialogue vol. 38, no. 4, December 2007

be – something ‘beyond’ calculable risk. In this context, uncertainty repre-
sents the incalculable fortunà that exogenously decides the right and wrong
of political planning. It represents a sudden qualitative change that ulti-
mately invalidates previous plans – even if they were conducted with the
best of intentions.

Under the conditions of the unknown unknowns, threats manifest themselves

in the form of disasters. A disaster happens instantly, without a warning, but
with major implications for the contours of the world polity. Such events
as the impact of meteorites, tsunamis and large terrorist attacks might be
analysed under this category. In particular, we think that people predomi-
nantly perceived the terror attacks of 11 September 2001 as a disaster – a one-
time thing that ‘happened’. Of course, to perceive terrorist attacks as a
disaster has important implications for the attribution of political responsi-
bility and for ‘regulatory demands’. To the extent to which terrorist attacks
are perceived as a disaster in the sense of the unknown unknowns, the FBI and
the secret services of the world are off the hook, as the possibility of ‘govern-
ing’, regulating and taming terrorism diminishes. Within the processes of
clarifying the circumstances of 9/11, there is a trend to present the incident as
a disaster and, at the same time, to regard the problem of future terrorist
attacks as a problem of ‘international risk management’. At this point, the
question of how unknown unknowns are transformed into known unknowns
becomes relevant. Equivalently, in probability theory, how might the cre-
ation of symmetry and probability occur? In the realm of security analysis,
this becomes the question: How are dangers – about which we simply do not
know – integrated into the political decision making process. In the 1990s, the
RAND Corporation carried out a study entitled ‘New Challenges for Defense
Planning’, which tried to provide an answer to this question. In one report,
Paul Davis (1994: 140) wrote that ‘most striking is the fact that we do not even
know who or what will constitute the most serious future threat’. To cope
with this problem, it would be mandatory, another RAND researcher wrote,
to break free from the ‘tyranny’ of plausible scenario planning. The decisive
step was to create ‘discontinuous scenarios . . . in which there is no plausible
audit trail or storyline from current events’ (Winnefeld, 1992: 15). These non-
standard scenarios were later termed the ‘wild cards’ and took on considerable
importance within the current US defence strategy. Here, the reactions
would not be ‘threat-based’, but they would be directed against potential
threats on ‘capability-based’ reasoning (Davis, 2002).

Example: Nuclear Terrorism

This is not the place to criticize the methodology of implausible scenario
planning. Instead, we simply want to point to one plain effect of such, which
was described by Carl Conetta & Charles Knight (1998: 36): ‘Conflict scen-

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

427

arios, both wild and tame, can gain more credibility in the telling than they
deserve.’ Cognitive studies speak here of an ‘Othello Effect’, as when false
accusations change Othello’s cognitive frame and ultimately lead him to
kill his beloved wife, Desdemona. The point is that even the most absurd
scenarios will gain in plausibility if a chain of potentialities changes cogni-
tion. They are thereby included in the realm of the possible, if not even the
probable: ‘Although the likelihood of the scenario dwindles with each step,
the residual impression is one of plausibility’ (Conetta & Knight, 1998: 38).
That this ‘Othello Effect’ is relevant for international relations is obvious: we
just need to look at the alleged connection between Saddam Hussein and Al-
Qaeda. What the US government tried to prove was disputed right from the
beginning. False evidence was repeatedly presented and repeatedly refuted.
But, that did not prevent the US government from attempting to present the
improbable yet possible connection of Iraq to terrorist networks, and the
improbable yet possible proliferation of an improbable yet possible nuclear
weapon to bin Laden as casus belli. As Donald Rumsfeld famously once said:
‘Absence of evidence is not evidence of absence’ – which is nothing else than
to admit that, under conditions of known unknowns and unknown unknowns,
different evidence criteria prevail. Unknown unknowns are not only hard to
identify, but also hard to refute.

Unknown Knowns and the Perfidies of Ignorance

To complete our discussion, a fourth form needs to be mentioned – the one
that Rumsfeld hid from the public. Here, the unknown is represented by
knowledge that is not known because it is not supposed to be known. This
not-to-be-wanted-to-be-known knowledge is a central category of political
decision processes, where information is systematically withheld or dis-
regarded as soon as it does not fit with operative concepts. This is not only a
problem in authoritarian regimes, as democracies are also vulnerable – if not
systematically exposed – to ‘group think’ and ‘cognitive biases’ (Festinger,
1957; Irving, 1982). The literature from organizational bias to ethnomethod-
ology, from cognitive science to post-modernism, all point to this tacit
knowledge that, nevertheless, is not only always present, but also a necessary
condition for social actions to occur. This kind of knowledge or non-
knowledge provides rhetorical force, and sometimes even leads to the strate-
gic use of the unsaid. The consequences are twofold: First, participation of
‘the public’ is less likely and may even be an unattainable ideal. Manipulation
of the public through one-sided or wrong information is one thing, but
(unconscious) self-manipulation is another. This is because the I-do-not-
want-to-know disposition leads to a self-chosen reduction of information and
political options and thus increases the danger of political miscalculations.
For example, because Larry Lindsay, economic adviser to US President

428

Security Dialogue vol. 38, no. 4, December 2007

George W. Bush, estimated the costs of the war and reconstruction of Iraq at
around US$100 billion, he was fired. Yet, President Bush now finds himself
in the awkward position of having to raise exactly three times

as much

through national and international political processes.

Ignorance

If threats emerge because of wrong decisions and non-knowledge, this is
often called ignorance – a concept heavily used in heterodox economics to
refer to the necessary limitedness of human knowledge. The almost complete
absence of any plan for the postwar situation in Iraq and the creation of a new
political order provides the best example of such. This became visible for the
first time at the beginning of April 2003, when insular troops of the Iraqi
Army shifted from conventional to guerrilla war tactics. Only two things
were surprising: first, that they waited for so long before doing so; second,
that the Americans were so surprised when they did. We are surely not in a
position to provide answers or a complete analysis of how to ‘win the peace’,
but it is quite astonishing to see the extent to which institutional and political
blind spots have led to such a wrong assessment of the situation. The politi-
cal ignorance was surely due to the political project of the current US
administration, who simply ignored warning voices and somehow thought
that the joy of freedom, as perceived and defined by the United States of
America, would bring about a political order by itself.

A second political ignorance lies in Donald Rumsfeld’s decision to be deaf

to the warnings of the Joint Chiefs of Staff, and to fight the war with the
inadequate number of soldiers that he deployed. In fact, given US tactics, it is
surprising that the Iraqi Army was unable to profit from the self-induced
vulnerabilities, in particular the securitization of US supply lines. The reason
for this blind spot was Rumsfeld’s determination to direct the Iraq War as an
example of modern warfare. Behind this was the idea that the so-called
Revolution in Military Affairs

allows wars to be fought with maximum

hi-tech equipment and minimum manpower. As far as that goes, Rumsfeld
might be right. But, the military victory was simply worthless as soon as it
became obvious that, in order to ‘win the peace’, far more than technology
was needed. What was needed most of all was a political plan and the per-
sonnel to secure the policing function. In addition, there is an institutional
ignorance on the part of the US military machinery – it seemingly was unable
to systemically learn from its experience in irregular warfare (Daase, 1999:
107–151). The USA has a remarkable short memory when it comes to how to
fight insurgencies. The reason for that might be the predominance of a
strategic culture that regards irregular warfare as a deviant form of war and

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

429

‘Projected Iraq War Costs Soar’, Washington Post, 27 April 2006.

See the discussion on the Rumsfeld Commission above.

glorifies the conventional, hi-tech-driven war. It might also have its causes in
the traditional contempt of the military for ‘political questions’, which seems
to be rooted in the history of the US military (Huntington, 1957). It is this dis-
regard of political questions that has its roots in the variety of knowledge –
and non-knowledge – structures that has made the US army appear ill-suited
for the current situation, as it is forced to relearn small-warfare tactics, which
results in exorbitant costs (Weighley, 1973). The current war-after-the-war
situation, we think, provides support for our argument that both knowledge
and non-knowledge are constitutive for the formulation of security policy.

Conclusion

This article has argued that non-knowledge is crucial for understanding
different logics of security policies and thus the understanding of social
processes. It particularly wants to convey the idea that political questions
need to address the form and type of non-knowledge that would actually
prevail in particular situations. Consequently, issues of non-knowledge
cannot be reduced to some ‘lack’ of knowledge that could be solved by look-
ing harder at the facts or processes of conjecture and refutation. Equivalently,
obvious insufficiencies associated with the current ‘fight against terrorism’
are not simply of an empirical nature. They cannot be solved by a simple
increase of knowledge, data and transparency, or better models. The ‘fight
against terrorism’ is characterized by the interplay of rather distinct logics
and forms of non-knowledge. The consequence for international relations
and security policy is that they cannot concentrate on ‘objective’ threats.
Rather, both theories and practices of international relations increasingly
need to come to terms with their own non-knowledge. This does not only
imply that one has to accept that threats can be over- or underestimated, or
even sometimes ignored. However, a position that would take non-
knowledge seriously needs to observe and incorporate its own contingency
(or probability), the contingency of its particular point of view, to examine
the form of non-knowledge that is used to frame threats.

This brings us back to Rumsfeld’s little poem. Rumsfeld’s neglect of the

fourth kind of non-knowledge points to the very reason why he crafted those
verses in the first place. It reveals the very episteme of the current ‘fight
against terrorism’: by hiding the things we do not want to know, Rumsfeld
hides ‘epistemic’ non-knowledge and therefore any inquiry into the histori-
cal or social context of terrorism. He is thereby in a position to present terror-
ism as a threat that can be deterred. However, as a consequence of
epistemological and ontological underpinnings, the ‘fight against terrorism’
is presented in terms of necessity and urgency – a fight where inaction

430

Security Dialogue vol. 38, no. 4, December 2007

would be irresponsible, in spite of the insufficient information at hand.
What is neglected is the contingency of one’s own position, and thus any self-
reflection that an acknowledgement of one’s own, but contingent, position
would require.

* Christopher Daase is Full Professor of International Relations at Ludwig-Maximilians-
Universität, Munich. Previously, he was Senior Lecturer at the University of Kent at
Canterbury and Director of the International Conflict Analysis Programme at the Brussels
School of International Studies. He has written extensively on IR theory and security
studies. Oliver Kessler is Assistant Professor for Political Science at the University of
Bielefeld. His most recent publications include ‘From Agent and Structures to Minds and
Bodies’ (in Journal of International Relations and Development) and (with Jan Helmig) ‘Space,
Boundaries and the Problem of Order: A View from Systems Theory’ (in Journal of
International Political Sociology).

References

Baldwin, David A., 1971. ‘Thinking about Threats’, Journal of Conflict Resolution 15(1):

71–78.

Beck, Ulrich, 1996. ‘World Risk Society as Cosmopolitan Society? Ecological Questions in

a Framework of Manufactured Uncertainties’, Theory, Culture and Society 13(4): 1–32.

Benenson, Frederick C., 1984. Probability, Objectivity and Evidence. London: Routledge &

Kegan Paul.

Bernstein, Peter L, 1998. Against the Gods: The Remarkable Story of Risk. New York: John Wiley.
Bonss, Wolfgang, 1991. ‘Unsicherheit und Gesellschaft. Argumente für eine soziologische

Risikoanalyse’ [Insecurity and Association: Arguments for a Sociological Risk
Analysis], Soziale Welt 42(2): 258–277.

Bonss, Wolfgang, 1995. Vom Risiko. Unsicherheit und Ungewissheit in der Moderne [About

Risk: Insecurity and Uncertainty in Modern Times]. Hamburg: Hamburger Edition HIS
Verlagsges.

Castel, Robert, 1991. ‘From Dangerousness to Risk’, in Gordon Burchell, Colin Gordon &

Peter Miller, eds, The Foucault Effect: Studies in Governmentability. London: Harvester
Wheatsheaf (281–298).

Cheney, Dick, 2003. ‘Cheney Cites “Risks of Inaction” with Iraq, CNN.com, 29 January

2003.

Cioffi-Revilla, Claudio, 1998. Politics and Uncertainty: Theory, Models and Application.

Cambridge: Cambridge University Press.

Cirincione, Joseph, 2000. ‘Assessing the Assessment: The 1999 National Intelligence

Estimate of the Ballistic Missile Threat’, Nonproliferation Review 7(1): 125–137.

Cohen, L. Jonathan, 1989. An Introduction to the Philosophy of Induction and Probability.

Oxford: Oxford University Press.

Cohen, Raymond, 1979. Threat Perception in International Crisis. Madison, WI: University of

Wisconsin Press.

Conetta, Carl & Charles Knight, 1998. ‘Inventing Threats’, Bulletin of the Atomic Scientists

54(2): 32–38.

Cooper, Richard N., 1996. ‘Emerging Missile Threats to North America During the Next

15 Years’, statement by the Chairman of the National Intelligence Council to the

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

431

House National Security Committee, 28 February 1996; available at http://www.
globalsecurity.org/space/library/congress/1996_h/h960228f.htm.

Daase, Christopher, 1999. Kleine Kriege – Große Wirkung. Wie unkonventionelle Kriegführung

die internationalen Beziehungen verändert [Small Wars, Big Impact: How Unconventional
War Transforms International Relations]. Baden-Baden: Nomos.

Daase, Christopher, 2002. ‘Internationale Risikopolitik. Ein Forschungsprogramm für den

sicherheitspolitischen Paradigmenwechsel’ [International Risk Policy: A Research
Programme for the Paradigm Change in Security Policy], in Christopher Daase,
Susanne Feske & Ingo Peters, eds, Internationale Risikopolitik [International Risk
Politics]. Baden-Baden: Nomos (9–35).

Davis, Paul K., 1994. New Challenges for Defense Planning: Rethinking How Much Is Enough.

Santa Monica, CA: RAND Corporation.

Davis, Paul K., 2002. Analytic Architecture for Capabilities-Based Planning, Mission-System

Analysis, and Transformation, RAND MR-1513-OSD. Santa Monica, CA: RAND
Corporation.

Davis, Paul K. & Brian Michael Jenkins, 2002. Deterrence and Influence in Counterterrorism:

A Component in the War on al Qaeda. Santa Monica, CA: RAND Corporation.

Douglas, Mary, 1992. Risk Acceptability According to the Social Sciences. London: Russell

Sage Foundation.

Douglas, Mary & Aaron Wildavsky, 1982. Risk and Blame: Essays in Cultural Theory.

London: Routledge.

Esposito, Elena, 2002. Soziales Vergessen: Formen und Medien des Gedächtnisses der Gesell-

schaft [Social Forgetting: Forms and Media of Social Memory]. Frankfurt: Suhrkamp.

Ewald, François, 1991. ‘Insurance and Risk’, in Gordon Burchell, Colin Gordon & Peter

Miller, eds, The Foucault Effect: Studies in Governmentability. London: Harvester
Wheatsheaf (197–210).

Falkenrath, Richard, 2001. ‘Analytic Models and Policy Prescription: Understanding

Recent Innovation in U.S. Counterterrorism’, Studies in Conflict and Terrorism 24(3):
159–181.

Festinger, Leon, 1957. A Theory of Cognitive Dissonance. Stanford, CA: Stanford University

Press.

Freedman, Lawrence, 1989. The Evolution of Nuclear Strategy, 2nd edn. Houndsmills:

Palgrave Macmillan.

George, Alexander L., 1993. Bridging the Gap: Theory and Practice in Foreign Policy.

Washington, DC: United States Institute of Peace Press.

George, Alexander L. & Richard Smoke, 1974. Deterrence in American Foreign Policy: Theory

and Practice. New York: Columbia University Press.

Good, Irving J., 1950. Probability and the Weighing of Evidence. London: Charles Griffin.
Good, Irving J., 1987. Good Thinking: The Foundations of Probability and Its Applications.

Minneapolis, MN: University of Minnesota Press.

Gotowicki, Stephen H., 1997. ‘Confronting Terrorism: New War Form or Mission

Impossible’, Military Review 77(3): 61–66.

Hacking, Ian, 1975. The Emergence of Probability. Cambridge: Cambridge University Press.
Hacking, Ian, 1983. Representing and Intervening. Cambridge: Cambridge University Press.
Hughes, Barry B., 1999. International Futures: Choices in the Face of Uncertainty. Boulder, CO:

Westview.

Huntington, Samuel P., 1957. The Soldier and the State: The Theory and Politics of

Civil–Military Relations. Cambridge, MA: Harvard University Press.

Irving, Janis, 1982. Groupthink: Psychological Studies of Policy Decisions and Fiascos, 2nd edn.

Boston, MA: Houghton Mifflin.

Jenkins, Brian M., 1999. ‘Testimony Before the Subcommittee on National Security,

432

Security Dialogue vol. 38, no. 4, December 2007

Veteran Affairs, and International Relations’, House Committee on Government
Reform, 106th Session of Congress, 20 October 1999.

Jervis, Robert, 1985. ‘Perceiving and Coping with Threat’, in Robert Jervis, Richard Ned

Lebow & Janice Gross Stein, Psychology and Deterrence. Baltimore, MD: Johns Hopkins
University Press (13–33).

Jones, Milo, 2004. ‘History in Foreign Affairs 1965–2000’, BSIS-Journal of International

Studies 1(1): 32–51.

Kadvany, John, 1996. ‘Taming Chance: Risk and the Quantification of Uncertainty’, Policy

Science 29(1): 1–27.

Kay, David, 2004. Dr David Kay’s Testimony to the Senate Armed Services Committee, US

Senate, 28 January; available at http://www.globalresearch.ca/articles/KAY401A.html.

Kessler, Oliver, 2007. Die Internationale Politische Ökonomie des Risikos [The International

Political Economy of Risk]. Wiesbaden: VS Verlag.

Keynes, John Maynard, 1937. ‘The General Theory of Employment’, Quarterly Journal of

Economics 51(2): 209–223.

Khong, Yuen Foong, 1992. Analogies at War: Korea, Munich, Dien Bien Phu, and the Vietnam

Decisions of 1965. Princeton, NJ: Princeton University Press.

Knight, Frank, 1921. Risk, Uncertainty, and Profit. Chicago, IL: Chicago University Press.
Knorr, Klaus, 1976. ‘Threat Perception’, in Klaus Knorr, ed., Historical Dimensions of

National Security Problems. Lawrence, KS: University Press of Kansas (78–119).

Koselleck, Reinhart, 1979. Vergangene Zukunft: zur Semantik geschichtlicher Zeiten [Past

Future: The Semantic of Historical Times]. Frankfurt: Suhrkamp.

Luhmann, Niklas, 1991. Soziologie des Risikos [Sociology of Risk]. Berlin: de Gruyter.
Luhmann, Niklas, 1997. Die Gesellschaft der Gesellschaft [The Society of Society]. Frankfurt:

Suhrkamp.

Münkler, Herfried, 2002. ‘Terrorismus als politisch-militärische Strategie’ [Terrorism as

Politico-Military Strategy], Merkur 56(1): 1–12.

National Intelligence Council (NIC), 1999. Foreign Missile Developments and the Ballistic

Missile Threat to the United States Through 2015; available at http://www.shaps.hawaii.
edu/security/nie99msl.html.

Pruitt, Dean G., 1965. ‘Definition of the Situation as a Determinant of International

Action’, in Herbert G. Kelman, ed., International Behavior. New York: Holt, Rinehart &
Winston (399–407).

Richardson, Robert C., 1989. ‘Risks and Implications of a Fading Threat’, Journal of Social,

Political and Economic Studies 14(4): 415–432.

Rumsfeld Commission, 1998. Report of the Commission To Assess the Ballistic Missile

Threat to the United States, July 15, 1998; available at http://fas.org/irp/threat/mis-
sile/rumsfeld/toc.htm.

Schelling, Thomas C., [1960] 1980. The Strategy of Conflict. Cambridge, MA: Harvard

University Press.

Seely, Hart, 2003. Pieces of Intelligence: The Existential Poetry of Donald H. Rumsfeld. New

York: Free Press.

Singer, J. David, 1958. ‘Threat-Perception and the Armament Tension Dilemma’, Journal of

Conflict Resolution 2(1): 90–105.

Smoke, Richard & Alexander L. George, 1973. ‘Theory for Policy in International Affairs’,

Policy Science 4: 387–413.

Snyder, Glenn, 1961. Deterrence and Defense. Princeton, NJ: Princeton University Press.
Stein, Janice Gross, 1989. ‘On Perceiving and Communicating Threats’, in Edward A.

Kolodziej, ed., Security and Arms Control: A Guide to National Policymaking, Vol. I.
London: Greenwood (25–47).

Walt, Stephen M., 1987. The Origins of Alliances. Ithaca, NY: Cornell University Press.

Christopher Daase & Oliver Kessler Uncertainty and the Political Construction of Danger

433

Weatherford, Roy, 1982. Philosophical Foundations of Probability Theory. London: Routledge

& Kegan Paul.

Weigley, Russell F., 1973. The American Way of War: A History of United States Military

Strategy and Policy. New York: Macmillan.

White, Jonathan R., 1998. Terrorism: An Introduction. Belmont, CA: Wadsworth.
Wills, Henry, ed., 2005. Estimating Terrorism Risk. Santa Monica, CA: RAND Corporation.
Winnefeld, James, 1992. The Post-Cold War Sizing Debate: Paradigms, Metaphors, and

Disconnects. Santa Monica, CA: RAND Corporation.

Zagare, Frank C. & Marc Kilgour, 2000. Perfect Deterrence. Cambridge: Cambridge

University Press.

434

Security Dialogue vol. 38, no. 4, December 2007