Complexity science and an understanding of the challenges associated with ‘wicked problems’ can provide insights into different disciplines, including nuclear weapons policy. Complexities can be found in the overall security environment, as well as within individual decision-makers themselves.
Experts define complex systems differently, depending on the disciplines to which they belong. Hence, there is no single definition of complex systems. This paper refers to complex systems as systems that are composed of a moderate-to-large number of mutually interacting sub-units (also referred to in the literature as factors, subsystems or agents), in which the interaction between these sub-units occurs in a non-linear way. This means that the components interact in such a way that the outcome is more than just the sum of the system’s parts.
There is interaction between the system and the environment. This points to the emergent behaviour in complex systems whereby the properties of the system cannot be explained through examining individual components per se, and whereby there is a degree of unpredictability within the system.
As the components are connected to each other, the actions of one sub-unit may have implications for those of another sub-unit, and the overall interaction of all sub-units may generate unpredictable behaviour. In some cases, complex systems also involve adaptation, whereby the components are either learning or modifying their behaviour through a feedback loop between each other and the environment. Within this characterization, some examples of a complex system would include a human brain, a genome, a colony of ants and social networks, as well as some engineering systems such as space shuttle programmes.,
Complex systems are often confused with chaotic systems. Although chaos theory led to the emergence of complexity theory, the two have different tenets. While chaos theory examines only ‘a few parameters and the dynamics of their values’, complex systems are ‘concerned with both the structure and the dynamics of systems and their interaction with their environment’. In a chaotic system, there is no cause-and-effect relationship.
It is important to differentiate between complex and complicated systems, in light of the fact that the terms are often (erroneously) used interchangeably. Complicated systems are predictable systems that operate in patterned ways. There are ‘known unknowns’ in complicated systems, i.e. the types of risks that the user may recognize beforehand but whose impact (in terms of the probabilities of an event) may not fully be understood. Thus, through rigorous analysis it is possible to observe a linear cause-and-effect relationship in a complicated system. A complicated problem could be managed through traditional systems engineering approaches, such as by applying best practices or standard operating procedures. Confusion of the terms ‘complex’ and ‘complicated’ often arises when a system has, for instance, both a complex design and a complicated structure. In a complex system, there are ‘unknown unknowns’, i.e. the types of incidents or errors that the user cannot or may not know beforehand. New forms of systems engineering rely on characteristics such as interdependent components, interacting feedback loops, open-system boundaries, self-organization, adaptation (e.g. learning ability) and emergence. These new characteristics render a system complex. In situations where there are unknown unknowns, decision-makers ‘can understand why things happen only in retrospect’.
The concept of ‘wicked problems’ was introduced in 1973 by Horst Rittel and Melvin Webber as a way of analysing problems involving multiple stakeholders with conflicting interests, whereby the stakeholders cannot agree on the problem. In a complex system, it is hard to judge the consequences of an action ahead of time: minor changes can lead to major impacts (also known as the ‘butterfly effect’) and major changes may have smaller impacts. It is likely that a ‘proposed solution to one segment of a wicked problem [could often lead] to unforeseen consequences in related segments’.
Wicked problems have a number of distinguishing characteristics, including the following:
- The causes of the problem can be explained in multiple ways, and the stakeholder’s choice of explanation determines the choice of resolution.
- There is no ‘true-or-false’ solution to a wicked problem: assessments of such solutions are instead expressed as ‘good’ or ‘bad’, because policy interventions are based on judgment rather than objective truth.
- There is no immediate or ultimate test of a solution.
- Every attempt to solve a wicked problem is important and has consequences for the system – thus, a ‘trial and error approach’ will not work. The attempt to fix the unintended consequences may also create a new set of wicked problems.
- Every wicked problem can be a symptom of another problem.
- The decision-maker has no right to be wrong, because of the number of potential consequences of their actions.
This study aims to provide insights through the lens of complexity science. The latter is defined as the study of different disciplines, examples from those disciplines, and wicked problems in order to revisit classical approaches to nuclear weapons policies. A focus on the framework of complexity and ‘wickedness’ that is inherent in nuclear weapons policy problems may offer new ideas for governance and new approaches for forward movement in the nuclear field.
Today’s international policymaking and relations rest on a three-dimensional complexity, similar to a game of three-dimensional chess: a) complexity in the issue area (i.e. nuclear weapons policy and new technologies); b) complexity in the overall security environment; and c) complexity in individual decisions of leaders and decision-makers. Added to this is the complexity that comes with internal bureaucratic politics in each country, whereby different organizations (e.g. ministries of foreign affairs, ministries of defence, etc.) working on nuclear weapons policy within the same country may view the problem and/or solution differently. Domestic political concerns are also a factor feeding into complexity, in that each political system – whether democratic or authoritarian – has to accommodate these concerns in some form within its policymaking.
In relation to nuclear weapons policy, nuances of opinion and different perspectives persist on the key questions, which might include the following: Do nuclear weapons act as an ultimate guarantee against large-scale war? What type of aggression do they deter? Are the risks associated with nuclear weapons postures negligible, or so high that they pose the threat of inadvertent catastrophe? In which areas are risks acceptable, in which are they unacceptable, and how does the threshold differ from one state to another? How should nuclear weapons possessors outside the Treaty on the Non-Proliferation of Nuclear Weapons (NPT) be addressed? Does the Treaty on the Prohibition of Nuclear Weapons (TPNW) complement or undermine the NPT? How does missile defence affect nuclear relationships? Is it better to have more actors in nuclear decision-making for checks and balances purposes, or fewer for speed and clarity? What type of measure can reduce misperception and misunderstanding in nuclear decision-making?
There are no definitive answers to these questions. Problems related to nuclear weapons policy, just like other policy issues, rely on political judgment, national interests, international commitments, technical capabilities and military strategies. Hence, nuclear weapons policy problems constitute a wicked problem.
Because of the complexity of wicked problems, there will never be agreement among experts as to what the problem is. This is liberating. For too long, experts and officials have all been trying to convince each other of what the problem is and what the solutions are. The answer is that nobody truly knows who is right or who is wrong – but complexity science may offer us new ways of thinking about the issue, provide insights and allow different communities to work together.
