Thesis-Chapter3-PolicyMaking.tex

\chapter{Strategy for Tackling Wicked Problems}
\label{chap-policy}

Before detailing the specific arguments used in the \ac{EA} debate, it is worth achieving a better understanding of
\acp{wicked-prob} in general. What are they, and why do the problems surrounding encryption, privacy, and \ac{EA} fall
into this category? What approaches to tackling \acp{wicked-prob} succeed and fail? How can we strategically confront
them and make real progress? This chapter seeks to answer these questions.


\section{Wicked Problems}

\Acp{wicked-prob} were previously introduced as pernicious and tricky issues that resist straightforward solutions.
This section analyzes the nature of wicked problems and approaches to tackling them.

\subsection{Characteristics}
\label{wicked-characteristics}

Rittel's categorization of \acp{wicked-prob} grew out of frustration with their resistance to traditional problem
solving methods. Since the Enlightenment, society has applied the scientific method to problems of every kind; the
sweeping application of scientific analysis has delivered reliable clean water, improved crop yields, shaped government
structures, and bestowed material wealth previously unimaginable. With these material problems largely solved in the
twentieth century, believers in the power of reason thought this progress would continue in the realm of public
planning. Policymaking would function by setting goals, identifying problems, evaluating alternatives, implementing
solutions, and analyzing outcomes in order to correct errors. Functioning as a continuous process, this approach was
primed to revolutionize governing the same way it did industry, agriculture, and economics---until it didn't. In the
context of what he describes as an ``anti-professional movement,'' Rittel explains how the scientific method has failed:

\begin{displayquote}
A great many barriers keep us from perfecting such a planning/governing system: theory is inadequate for decent
forecasting; our intelligence is insufficient to our tasks; plurality of objectives held by pluralities of politics
makes it impossible to pursue unitary aims; and so on. The difficulties attached to rationality are tenacious, and we
have so far been unable to get untangled from their web. This is partly because the classical paradigm of science and
engineering---the paradigm that has underlain modern professionalism---is not applicable to the problems of open
societal systems. One reason the publics have been attacking the social professions, we believe, is that the cognitive
and occupational styles of the professions---mimicking the cognitive style of science and the occupational style of
engineering---have just not worked on a wide array of social problems. The lay customers are complaining because
planners and other professionals have not succeeded in solving the problems they claimed they could solve. We shall want
to suggest that the social professions were misled somewhere along the line into assuming they could be applied
scientists---that they could solve problems in the ways scientists can solve their sorts of problems. The error has been
a serious one. \cite{rittel_dilemmas_1973}
\end{displayquote}

When applied to social problems, the prescribed method---here defined as setting goals, identifying problems, evaluating
alternatives, implementing solutions, and analyzing outcomes---fails at every step. In the U.S., a nation composed of
many varying cultures and subcultures and politically dominated by two mutually hostile parties, agreeing on
goals is a challenge in itself. When goals are set, we often discover that we are contending with \acp{wicked-prob} that
defy the process at each remaining step. Rittel provides a list of ten characteristics of problems in this category
\cite{rittel_dilemmas_1973}:

\begin{displayquote}
  \begin{enumerate}
    \item There is no definitive formulation of a wicked problem [and each formulation presupposes a solution].
    \item Wicked problems have no stopping rule.
    \item Solutions to wicked problems are not true-or-false, but good-or-bad.
    \item There is no immediate and no ultimate test of a solution to a wicked problem.
    \item Every solution to a wicked problem is a ``one-shot operation''; because there is no opportunity to learn by
          trial-and-error, every attempt counts significantly.
    \item Wicked problems do not have an enumerable (or an exhaustively describable) set of potential solutions, nor is
          there a well-described set of permissible operations that may be incorporated into the plan.
    \item Every wicked problem is essentially unique.
    \item Every wicked problem can be considered to be a symptom of another problem.
    \item The existence of a discrepancy representing a wicked problem can be explained in numerous ways. The choice of
          explanation determines the nature of the problem's resolution.
    \item The [decision maker] has no right to be wrong [i.e., they are liable for the consequences of their decisions].
  \end{enumerate}
\end{displayquote}

In summary, (1) the \acp{wicked-prob} are impossible to definitively identify, (2) potential solutions are impossible to
definitively evaluate, (3) every action or inaction has permanent effects, and (4) solutions are importable from one
problem to another. The policy version of the scientific method cannot be used under these conditions. The ``wicked''
problem's counterpart is the ``tame'' problem. ``Tame'' does not imply easy; it only means that the scientific approach
will be effective.

Several of \acp{wicked-prob}' characteristics are results of the fact that we have no accurate predictive model of the
world and human behavior. This is self-evident. The more important insight is one step removed: precisely because there
is an inexhaustible set of potential solutions, the problem definition---that is, the set of information required to
produce a solution---is not self-contained. Each proposed solution demands that new research and context be fed back
into the problem definition. This feedback from proposal to definition violates the linearity of the traditional
approach. One cannot reason from problem statement to proposals to a solution; instead, one is forced to constantly
refine the problem statement based on the content of proposals themselves.

A 2018 study by the Australian government uses climate change as an example of a \ac{wicked-prob}
\cite{commission_tackling_2018}. Say the problem is initially formulated as long-term change to the environment caused
by the effects of accumulating greenhouse gasses. Responses typically take one of three broad forms
\cite{commission_tackling_2018}. First, profligate behavior in consumeristic societies must be reigned in at a local and
personal level. Second, global inter-governmental coordination is the only solution, as individual changes will have no
impact. Third, the situation is overblown by idealists and power-mongers, and technological progress and adaptive
markets will handle any negative effects that come to pass. Which response is correct? It is impossible to tell by
evaluating the problem statement. How much do individual choices contribute to greenhouse gas emissions? How effective
are international accords? How costly will the changes be, and how capable is technology to respond? Evaluating the
validity of each proposal requires updates to the problem statement itself.

% Elements of the problem may be tame---for example, one can develop a new renewable energy widget and another can
% independently confirm its efficiency. Other elements of the problem are wicked---for example, one cannot predict the
% side effects of the deploying the widget, and there is now way to know how effort spent there could have been used
% elsewhere. \Acp{wicked-prob} always consist of subproblems, which may be tame or wicked in and of themselves

\Acp{wicked-prob} always consist of subproblems, which may be tame or wicked themselves. For example, climate change has
many subproblems. Developing clean energy technology is a tame subproblem; implementing any solution in a way that
doesn't leave vast swaths of people behind is a wicked subproblem. \Acp{wicked-prob} quickly grow in complexity with the
number of subproblems that comprise them. Truly imposing \acp{wicked-prob} are composed of a tangle of contributing
factors, and despite the presence of tame elements, they pose a large number of difficult challenges.

% ... WPs quickly grow in complexity with the number of subproblems that comprise them. Refining the problem statement
% and combining tame solutions with wicked subproblem mitigations is the truly ``wicked'' part.

% Elements of the problem may be tame---for example, one can independently confirm that reducing global carbon emissions
% by a certain amount will result in a certain difference in response with a certain statistical confidence. Other
% elements of the problem are wicked---for example, one cannot verify the effectiveness of a plan ahead of time, port a
% solution from a different domain, or definitively evaluate its performance after the fact. Refining the problem
% statement and combining the tame elements into a cohesive solution is the ``wicked'' part.

% Elements of the problem may be tame---one can independently confirm that reducing global carbon emissions by a certain
% amount will result in a certain difference in response with a certain statistical confidence. \Acp{wicked-prob} always
% consist of subproblems, which may be tame or wicked in and of themselves. Refining the problem statement and combining
% the tame elements in a cohesive solution is the ``wicked'' part. Despite certain tame elements, one cannot verify the
% effectiveness of a plan ahead of time, port a solution from a different domain, or even definitively evaluate its
% performance after the fact.

\subsection{Encryption and EA as a Wicked Problem}

Dogged by concerns over privacy, security, safety, and trust, \ac{encryption} and the presupposed solution of \ac{EA} is
a \ac{wicked-prob}. It has each of the characteristics from Rittel's list above:

\begin{enumerate}
  \item There is no formulation that encapsulates the problem of encryption's interplay with privacy, security,
        safety, and trust.
  \item Balancing each value in the face of constantly evolving technology is a never-ending cycle.
  \item \ac{EA} or other proposals cannot definitively solve the problem.
  \item \ac{EA} or other proposals cannot be objectively tested.
  \item Every policy implementation has irreversible effects.
  \item There is an inexhaustible set of potential solutions to the problem.
  \item Solutions from other domains do not apply directly to this problem.
  \item The need for \ac{EA} or some alternative is a symptom of differing values, rapid technological change, and
        criminal behavior.
  \item The problem can be framed many ways: as insufficient investigatory access to digital data (presupposing \ac{EA}
        as the solution), outdated cyberlaw (presupposing legal hacking or compelled password disclosure as solutions),
        and more.
  \item The decisions of regulators and technologists have real impacts in the world today.
\end{enumerate}

Cybersecurity law scholar Alan Rozenshtein argues at length for treating \ac{encryption} and \ac{EA} as a
\ac{wicked-prob} \cite{rozenshtein_wicked_2018}. He divides the root issues into three categories. First, there is
disagreement on what the goals should be (and even basic premises). Sides do not agree on how much to value competing
notions of security. Regarding the basic facts, sides do not agree on whether encryption is hiding so much evidence that
law enforcement is ``going dark,'' as one side puts it, or whether technological change overall has created a ``golden
age for surveillance,'' as their opponents argue. Second, information is ``uncertain and diffuse.'' Comprehensive data
regarding encryption's effect on investigations is unavailable. Although the consensus is that we are not presently
capable of acceptably secure \ac{EA}, that consensus could change, especially considering that \ac{EA} as a field is
under-researched. Third, the problem cannot be definitively solved. Evolving values and technology mean that this policy
area is always up for renegotiation.

\Ac{encryption} technology is particularly sensitive to the irreversible effects of policy implementation. Technology
deployments have long tails and attackers have the ability to record and store data for later analysis. In one
investigation, Australian police cracked a cold case based on evidence acquired from a mobile device that they cracked
five years after it was seized \cite{evans_arrests_2020}. In this case, it was lawful authorities that benefitted from a
vulnerability. However, if miscalculated \ac{EA} mandates result in vulnerabilities such as this, attackers will
benefit, not just law enforcement.

Rozenshtein reflects on the \ac{wicked-prob} diagnosis optimistically:

\begin{displayquote}
Recognizing that something is a wicked problem is not an admission of its insolubility; rather, it’s just a realistic
appreciation of its challenges. Progress on difficult social problems reflects, almost by definition, progress on wicked
problems, whether economic inequality, environmental degradation, or government access to data. Progress can be made,
but it first requires a clear-eyed appreciation of the nature of the problem and the nature of its challenges.
\cite{rozenshtein_wicked_2018}
\end{displayquote}

Reality must be accepted before it can be dealt with. \cite{baker_2019}. We have by now embraced the reality of
\acp{wicked-prob}. The next two sections investigate strategies for dealing with this reality.


\section{Failure of Current Policymaking Approaches}

This section describes two common policymaking approaches, the \ac{classical-method} and \ac{incrementalism}.

\subsection{The Classical Analytic Method}

The \ac{classical-method} \cite{feeley_judicial_2000} (also known as ``the modern-classical model of planning''
\cite{rittel_dilemmas_1973}, ``the rational-comprehensive method'' \cite{lindblom_muddling_1959}, ``traditional policy
analysis'' \cite{rozenshtein_wicked_2018}, or ``linear thinking'' \cite{commission_tackling_2018}) has already been
introduced. It is the reason-based method that functions by setting goals, identifying problems, evaluating
alternatives, implementing solutions, and analyzing outcomes in order to correct errors. \myfig{fig-classical-method}
illustrates this approach in the context of encryption and \ac{EA}. It has a purely linear flow from problem to solution
except for the ``refinement'' step, in which the method analyzes outcomes and corrects errors. However, it is important
to note that refinement represents a reinforcement, as opposed to a reassessment, of the chosen solution.

\begin{figure}[h]
  \centering\CaptionFontSize
  \includegraphics[width=\linewidth]{dfds/build/OODA-classical.png}
  \caption{The Classical Analytic Method}
  \label{fig-classical-method}
\end{figure}

The shortcomings of the \ac{classical-method} were discussed in \mysec{wicked-characteristics}. It fails due to
disagreement over goals, the dependence of the problem definition on the potential solutions generated (violating the
linear flow), the inability to evaluate alternatives, and the lack of a definitive stopping rule.

\subsection{Incrementalism}

\Ac{incrementalism} is an intuitive and iterative approach posed as an alternative to the \ac{classical-method}.
\Ac{incrementalism} as a policymaking strategy is often referred to as ``\ac{muddling-through}'' from political
scientist Charles Lindblom's classic paper defining and defending the approach \cite{lindblom_muddling_1959}. Written
before Rittel developed the idea of ``\acp{wicked-prob},'' Lindblom nonetheless identified many of the same shortcomings
of the classical method and sought to formalize the process policymakers were already often using.

Lindblom's process of ``\ac{muddling-through}'' operates by taking successive steps chosen through comparative analysis.
The alternatives selected for comparison must be defined relative to the status quo and must be close enough to one
another that they can be analyzed on the margin. This is done due to (a) practical necessity, due to the inability to
predict policy outcomes, and (b) out of political realism, as non-incremental changes are usually politically impossible
to impose in a democratic system \cite{lindblom_muddling_1959}. The formality of the process varies; policymakers may
use this method consciously, with considerable comparative analysis, or unconsciously, led by intuition.
\myfig{fig-muddling-through} illustrates this method.

%Footnote: see Lindblom's bracing quote: ``Party behavior is in turn rooted in public attitudes, and political theorists
%cannot conceive of democracy's surviving in the United States in the absence of fundamental agreement on potentially
%disruptive issues, with consequent limitation of policy debates to relatively small differences in policy.''

\begin{figure}[h]
  \centering\CaptionFontSize
  \includegraphics[width=0.55\linewidth]{dfds/build/OODA-muddling.png}
  \caption{Incrementalism}
  \label{fig-muddling-through}
\end{figure}

\Ac{incrementalism} has several advantages over the \ac{classical-method} for handling \acp{wicked-prob}. It is rooted
in realism about the limits of rational analysis. It accepts that the problem will not be conclusively solved. Instead,
it emphasizes iteration: ``Policy is not made once and for all; it is made and re-made endlessly. Policy-making is a
process of successive approximation to some desired objectives in which what is desired itself continues to change under
reconsideration'' \cite{lindblom_muddling_1959}. Most importantly, it eschews the linear flow from problem definition to
alternatives analysis out of respect that the problem definition is not self-contained. Lindblom argues that policy ends
and means are interlinked, eventually concluding:

\begin{displayquote}
As to whether the attempt to clarify objectives in advance of policy selection is more or less rational than the close
intertwining of marginal evaluation and empirical analysis, the principal difference established is that for complex
[i.e., wicked] problems the first is impossible and irrelevant, and the second is both possible and relevant. The second
is possible because the administrator need not try to analyze any values except the values by which alternative policies
differ and need not be concerned with them except as they differ marginally. His need for information on values or
objectives is drastically reduced as compared with the root [i.e., classical analytic] method; and his capacity for
grasping, comprehending, and relating values to one another is not strained beyond the breaking point.
\cite{lindblom_muddling_1959}
\end{displayquote}

Despite its strengths, \ac{incrementalism} also has weaknesses in its ability to address \acp{wicked-prob}. Its main
weakness is the absence of high level strategic analysis. Incrementalism is incapable of drastic change, which is
sometimes necessary. By Lindblom's admission, it lacks a safeguard for consideration of all relevant values and may
``overlook excellent policies for no other reason than that they are not suggested by the chain of successive policy
steps leading up to the present'' \cite{lindblom_muddling_1959}. Analysis based only on the present status quo can
result in messy policy that pleases no one---``As Lindblom's sobriquet suggests, it often [leads] to a considerable
muddle'' \cite{feeley_judicial_2000}. Application of the \acl{CFAA} is one such muddle \cite{wolff_computer_2016}.

Unfortunately, one cannot simply add high level strategic analysis to the incrementalist method by including the past in
its analysis. \Ac{incrementalism} relies on simplifying analysis by limiting it to marginal differences to a given
baseline. If one tries to be a ``strategic incrementalist'' by looking into the past, they still have to choose a
baseline from which to perform analysis. This approach is susceptible to two weaknesses in baseline-based reasoning that
Rozenshtein describes in his analysis of \ac{EA} as a \ac{wicked-prob} \cite{rozenshtein_wicked_2018}. First, the choice
of baseline is arbitrary, yet heavily colors analysis:

\begin{displayquote}
Both the government and its critics have operated from the status-quo baseline, though from opposite directions. For the
government, the relevant baseline is recent history---specifically, right before companies like Apple and WhatsApp
encrypted their products. From this baseline, the government's ability to surveil has diminished. For critics of
government surveillance, the relevant baseline is the pre-digital age, before smartphones and social media vastly
expanded the government's surveillance capabilities. From this baseline, the technological changes underlying the
``going dark'' problem are mere blips on the otherwise rocketing growth of the surveillance state.
\cite{rozenshtein_wicked_2018}
\end{displayquote}

Second, unlike legal baselines, policy baselines do not carry normative force. Constant changes in the underlying
situation mean that even optimal policy in the past is not necessarily desirable in the present. Due to these
weaknesses, applying incrementalist methods at the strategic level does not work.

A final weakness in \ac{incrementalism} is its assumption of basic agreement and political stability. The method works
by limiting analysis to marginal comparisons of broadly similar and familiar proposals. Proposals that differ widely
from one another or the status quo are considered irrelevant because the debating parties both share the same general
goals and lack the ability to unilaterally impose their will. Rather a symptom of present circumstances (see
\mysec{sec-history-current}) than an inherent weakness in the incrementalist approach, both of these assumptions are
incorrect.


\subsection{Lessons}
\label{sec-policy-lessons}

The \ac{classical-method} and \ac{incrementalism} are not the only styles of policymaking. They represent perhaps two
extremes on a spectrum of rational planning and intuitive ``muddling.'' Both have strengths and weaknesses. One may
intuit that a reasonable strategy is the selective use of both approaches according to the situation, and indeed this
has been formally suggested \cite{etzioni_scanning_1967}. However, even a combination of these methods does not suit all
classes of wicked problems. Is a problem tame? Use the \ac{classical-method}. Is it wicked, but strategically under
control and relatively non-controversial? Use \ac{incrementalism}. What about when it is it wicked, lacks a strategic
response, and is highly controversial?

\Acp{wicked-prob} were earlier described as a tangle of contributing subproblems. Using this characterization, dealing
with a \ac{wicked-prob} requires disentangling the web, identifying the tame subproblems, and developing agreeable
strategies for the irreducibly wicked subproblems. Once that is done, we can rationally analyze and increment our way to
resolution. There is no handbook for how to do this, but several sources offer advice.

\newcommand{\wickedtipsstart}{\begin{itemize}}
\newcommand{\wickedtip}[2]{ % Name, citation, description
  \item \textbf{#1} \nopagebreak

  \vspace{0.5\baselineskip} \parbox{\linewidth}{#2} \vspace{0.5\baselineskip}
}
\newcommand{\wickedtipsend}{\end{itemize}}

\wickedtipsstart

  \wickedtip{Reject Easy Answers}{

Easy answers, or any solutions that artificially tame the problem, will not bring the matter under control. While by
definition there is no solution that will truly solve a \ac{wicked-prob}, easy answers deliberately emphasize one value
to the exclusion of others. Because they neglect the root issues, solutions based on easy answers produce unintended
consequences and chaos in those neglected areas \cite{commission_tackling_2018}.

}

  \wickedtip{Bring Everyone to the Table}{

Including every relevant group is important for two reasons. First, because information is ``uncertain and diffuse,''
generating an accurate problem statement requires diverse input \cite{rozenshtein_wicked_2018}. Second, for
\ac{incrementalism} to work on the irreducibly wicked roots of the problem, there needs to be some degree of consensus
on overall strategy. Consensus building is not easy among groups with differing values and priorities, but it is
impossible without each group being represented.

% TODO: Prof. Engelsma:
% - A big part of the "solution" is getting people together
% - My thesis seems aimed at tech people to get them to understand and judge threats
% - How can I take this accessible and bridge the gap for policymakers?

}

  \wickedtip{Unite Problem Definition and Analysis Steps}{

The failures of the \ac{classical-method} and \ac{incrementalism} reveal that we must be able to think strategically
while respecting the non-linear nature of \acp{wicked-prob}. One must use the high level, holistic view of the classical
approach while intertwining the problem definition and analysis steps as in the incrementalist approach. In practice,
this means that the collective understanding of the problem and potential solutions must co-evolve. As Rittel puts it,
``The systems-approach `of the first generation' [i.e., classical analysis] is inadequate for dealing with
wicked-problems. Approaches of the `second generation' should be based on a model of planning as an argumentative
process in the course of which an image of the problem and of the solution emerges gradually among the participants, as
a product of incessant judgment, subjected to critical argument'' \cite{rittel_dilemmas_1973}.

}

  \wickedtip{Embrace Flexible, Risk-Based Solutions}{

\Acp{wicked-prob}' potential solutions cannot be comprehensively evaluated before or even after implementation and each
action (or decision not to act) has irreversible effects. Proposals must therefore be agile. All decisions involve
unknowns, but risk- and uncertainty-management strategies can optimize the expected outcome and maximize the worst
outcome \cite{sunstein_beyond_2015}.

}

  \wickedtip{Focus Discussion around Concrete Proposals}{

Focusing on concrete proposals follows partially from the previous lesson---problem definition and analysis are combined
precisely because the problem definition depends on the nature of proposed solutions. But this advice deserves emphasis
for another reason: consensus is easier to achieve for concrete proposals. Debate in the abstract can rage endlessly, as
groups are bound to disagree due to their conflicting values and priorities. However, Lindblom writes encouragingly
about ``the ease with which individuals of different ideologies often can agree on concrete policy'' in an example about
congressional compromise. He goes on to say, ``Labor mediators report a similar phenomenon: the contestants cannot agree
on criteria for settling their disputes but can agree on specific proposals. Similarly, when one administrator's
objective turns out to be another's means, they often can agree on policy'' \cite{lindblom_muddling_1959}.

}

\wickedtipsend


\section{Proposal: The OODA Loop for Wicked Problems}

The discussion above describes the inability of well-established methods of policymaking to address \acp{wicked-prob}
and lessons for shaping a better method. Here, I propose a modified \acs{OODAloop} as an alternative policymaking model.

The \ac{OODAloop} was developed by Air Force Colonel John Boyd as a description of a successful strategy for countering
opponents in real-time combat \cite{angerman_2004}. Boyd created it in the context of military strategy, but its ideas
have penetrated many other sectors, including cybersecurity, where it serves as a model for structuring incident
response \cite{schneier_future_2014}. The model emphasizes fast cycle times and ``getting inside'' your opponent's loop
as a means of overcoming raw power with speed and agility. It is illustrated in \myfig{fig-ooda-loop}.

\begin{figure}[h]
    \centering\CaptionFontSize
    \includegraphics[width=\linewidth]{dfds/build/OODA-vanilla.png}
    \caption{The OODA Loop}
    \label{fig-ooda-loop}
\end{figure}

The \ac{OODAloop} serves well as a model for contending with \acp{wicked-prob} because it has the correct fundamental
structure. It corresponds closely to the \ac{classical-method} steps of problem definition (orientation), ideation and
analysis (orientation), decision, and action; however, it also emphasizes \ac{incrementalism}'s iteration and feedback.
Additionally, while policymaking does not share the pace of real-time combat, this metaphor holds true in other ways. In
both cases, one faces an unpredictable opponent in a dynamic environment in which every action counts. We cannot
overpower \acp{wicked-prob} through force of reason, but perhaps we can through feedback and agility.

The OODA Loop does need one modification in order to suit wicked problems. Because the problem definition step and
analysis step depend on one another, they must be joined. This is Rittel's ``argumentative process in the course of
which an image of the problem and of the solution emerges gradually'' \cite{rittel_dilemmas_1973}, illustrated in
\myfig{fig-policy-ooda-loop} as a cycle of ``collaborative debate'' between observation and orientation. The modified
\ac{OODAloop} incorporates each of the lessons from \mysec{sec-policy-lessons}.

\begin{figure}[h]
    \centering\CaptionFontSize
    \includegraphics[width=\linewidth]{dfds/build/OODA-loop.png}
    \caption{The OODA Loop for Wicked Problems}
    \label{fig-policy-ooda-loop}
\end{figure}


\section{Summary}

Traditional policymaking typically follows either (a) a rational approach rooted in the scientific method, aiming to be
thorough and complete, or (b) an intuitional approach rooted in evolutionary trial and error, humbly aiming only to take
steps in the right direction. Both approaches have strengths, but neither is suited to the complex and controversial
nature of \acp{wicked-prob}. The strategy proposed here relies on the goodwill participation of all parties who engage
in ongoing research and debate that rejects easy answers, encourages flexible risk-based solutions, and crystallizes
discussion around specific proposals. The result of a single iteration is a bit more clarity and a small step forward.
The result of many iterations is the breakdown of the problem into subproblems---some tame and inevitably some still
wicked. The tame subproblems can be addressed with the \ac{classical-method}, and the wicked subproblems, by now
restrained under a sound and agreeable strategy, can be addressed with \ac{incrementalism}.

This process is demonstrated in \myfig{fig-policy-ooda-process}. An out-of-control problem (represented by the large
storm) is confronted in this diagram. Several iterations of the \ac{OODAloop} gradually diminish its size. This is
achieved by spinning out tame problems (represented by puzzle pieces) which are addressed with classical analysis and
wicked but restrained problems (represented by small storms) which are controlled by \ac{incrementalism}.

\begin{figure}[h]
  \centering\CaptionFontSize
  \includegraphics[width=\linewidth]{dfds/build/OODA-policy-process.png}
  \caption{Using the \acs{OODAloop} to Tackle Wicked Problems}
  \label{fig-policy-ooda-process}
\end{figure}

It is important to note that the same \ac{wicked-prob} is never faced twice, since each one is by definition, unique.
However, changes in technology, culture, and current affairs may render the current strategy insufficient and require
the process to start again. It has been said that history does not repeat itself, but it rhymes. \Acp{wicked-prob} are
the same. Past solutions cannot be used in the future, but lessons learned can be.

In the context of \ac{encryption} and \ac{EA}, this means that the outcome of \ac{the-cw1} doesn't hold precedential
power over the debate in the second. It means that it is legitimate to re-raise questions about the role technology
plays in society. Hard-line rhetoric doesn't help \cite{ruiz_there_2018} \cite{geller_2019}, but sincere appeals do
\cite{abelson_2015} \cite{intl_2020} \cite{rozenshtein_2019}. Collaborative efforts involving government, technical, and
civil liberties representatives are even better \cite{committee_decrypting_2018} \cite{group_2019}. Specific proposals
around which groups can center discussion are also necessary. This is true whether they specify a particular form of
\ac{EA} or (perhaps especially) if they offer an alternative \cite{kerr_encryption_2017} \cite{wright_crypto_2018}
\cite{phan_key_2017}.

I apply this strategy in the remainder of the thesis. \mychap{chap-arguments} documents the debate in depth and
\mychap{chap-threatmodel} defines a threat model against which I analyze a specific \ac{EA} proposal.