Mental models of cities and their relevance to urban resilience

Magnus Moglia, Fabio Boschetti, and Nicky Grigg
CSIRO, AUS

Iain Walker
University of Canberra, AUS

ABSTRACT

We believe that cities are important for humans as essential forms of social organization in contemporary human life. Currently, the integrity of cities as enduring systems faces many challenges—'exogenous' factors such as unsustainable consumption of energy and other resources and 'endogenous' factors such as 'livability' and the 'human scale' of cities. Therefore we must work to ensure their future, hence the emerging importance of the concept of resilience. But how do we ensure the future of cities? Current slow, de-centralized and business-as-usual urban development is problematic. Instead, a planned approach to urban development is necessary, but how do we plan for cities to be resilient? Planning must inevitably rest on an understanding of how a city functions, and this leads us to thinking of developing mental or computational models of cities. In this paper we explore a number of mental models of cities, which could form the basis for directed urban planning. We identify three types of urban models, urban- state models, urban-learning models, and urban-systems models. Furthermore, we argue that all the current urban models are piecemeal and/or impractical and either do not adequately consider the complexity of the city or are not suitable for the interface with governance, We suggest that the best way forward is to embed multiple urban models within an adaptive governance framework, thereby providing a way for urban decision makers and planning organizations to better handle the complexity of their cities. To enable this, further work is required to identify suitable urban systems archetypes.

Introduction

Managing a complex adaptive system is a challenge encountered in many domains. We need to manage firms, factories, infrastructure, natural resources, economies and nations. For some systems, like firms and factories, the purpose of the management may appear well defined (maximizing profits or production), although a more careful analysis often reveals that it may also involve proper accounting of many more factors, like balancing short vs long-term profits, power tensions within the organization, future expectations, etc. For other systems, such trade-offs are more explicit. For example, different people see different purposes for managing a natural resource (profit maximization, ecosystem service provision, employment, environmental conservation, respect of cultural values, etc.), some of which may be irreconcilable.

The complexity inherent in managing a complex adaptive system comes to full display when we consider how to manage a city. It is so because not only is the purpose of urban governance contested, but also the very essence and functions of a city are understood differently by different parties because the multi-faceted nature of a city imposes inevitably narrow experiences to its inhabitants.

It is therefore unsurprising that the governance of a city needs to be thought of in terms of systems properties, and this is why urban resilience is an increasing popular and desired goal of urban governance. We are particularly interested in the notion of urban resilience and will explore this concept further. In this work we adopt the definition of urban emergence proposed in (Meerow & Newell, 2016) which we included in Table 1. Meerow and colleagues (Meerow & Newell, 2016) have also undertaken the most comprehensive review of urban resilience notions so far and has uncovered a set of related conceptual tensions inherent in the literature on urban resilience:

• It is often unclear where the boundaries of 'the urban' start and finish, both in spatial and temporal terms; as well as in terms of social and political terms. Modern cities interact with their region and are closely intertwined with global processes. Furthermore, a city can be thought of as both a socio-ecological and a socio-technical system. We argue that a city system needs to be thought of in the context of multiple scales, and simultaneously socio-ecological and sociotechnical.
• Steady state vs non-equilibrium, i.e., much of the urban literature assumes a steady state, yet cities undergo constant change. In terms of resilience, it would seem that a more useful goal is to try to maintain the desired function, within an ever- changing system.
• Resilience as a negative, neutral or positive concept. As resilience refers to the capacity to maintain and/or return to a 'normal' state, its value judgment depends on whether that normal state is desirable. It needs to be acknowledged that the desired state is contested, and urban resilience refers to the ability to maintain a 'desired state', even if this state is contested.
• The range of pathway to change vary in literature, i.e., persistence, transition or transformation, often relating to the varying degree and mechanisms of required change and the speed of change. We argue that due to the pressures of the 21^st century, cities need to be completely reimagined and transformed in order to maintain the functions required for the future viability of humanity. Luckily, there is significant potential for technological and social reform progress to allow this to happen.
• Specific vs general resilience. Should the city be resilient to specific or general threats? The specific threats may relate to natural disasters, drought or resource shortages. The general threats, however, act at a system level, are unpredictable and hard to prepare for and would warrant a general adaptive capacity and flexibility to change. We argue that much of the urban adaptive capacity to both specific and general threats is inherent in the governance systems, and therefore the focus should be on urban governance processes.
• The timescales that people focus on vary depending on the issue under consideration, i.e., natural disasters play out over short timeframes and climate change plays out over the longer timeframe. Future generations are occasionally considered, but perhaps, more importantly, the question of timeframe neglects the anticipatory capacity of urban governance systems as a critical aspect of urban resilience. We argue that the time frame is both short term and long term and involves future communities.

As mentioned, the notion of urban resilience is inevitably contested. The desired functions of a city change depending on what people understand a city to be (a built environment? A large group of people? A political entity?), how it functions (it evolves as a whole? It has a life in itself? It responds to social development? It responds to politicians and businesses?) What role it plays (it generates economic growth? it produces pollution? it fosters culture and innovation?). A city can be different things to different people in different circumstances. Notwithstanding the contested nature of urban resilience, we believe the definition included in Table 1 provides a good starting point in addressing the conceptual tensions. However, we believe two additional points are needed to place the definition firmly in the context of the 21^st century:

• A significant urban transformation is necessary to ensure that the global environmental footprint helps the planet to be within a safe operating space (McPhearson et al., 2016; Webb et al., 2018).
• An anticipatory capacity is necessary in order to foresee future challenges, and to identify and devise strategies to avoid significant hardship (Moglia et al., 2018).

These points also lead us to the belief that for urban transformation to occur at the necessary speed, cities need to be managed in a deliberate manner (Webb et al., 2018).

While, subjectively, urban resilience is a positive notion only when aligned with desired goals and functions that are embedded within it, let us here assume that a city governance system has defined a goal and function that most of its current, future and past inhabitants could agree with. What then? How does a city ensure this goal and function into the future?

To understand how a city could adapt and change, either in anticipation of hardship, or in response to hardship, or simply to improve the ability of the city to achieve desired functions, urban managers need to choose how to act. In systems terms this requires the identification of the system leverage points (Meadows, 2008) and in turns a representation, be it mental models or computer models, of how cities function (Bai et al., 2016). Definitions of mental and computational models, and how they differ is provided in Table 1.

In this article we report on reflections from our experience in selecting a suitable computer model of urban dynamics for a project aimed at studying the resilience of large Australian cities (Moglia et al., 2018). We also report on reviews and surveys of mental and computational models of urban systems, and how they align with the need of governance systems to support urban resilience.

Table 1: Glossary of terms used in this work

Three types of urban mental models

Addressing urban resilience requires that we first analyze how we understand a city. Most people are unlikely ever to think deliberately about what they understand a city to be. But when asked, they are likely to articulate some sort of understanding of a city either as a metaphor or as a more or less well-defined mental model. We refer to these as 'urban metaphors' and 'urban mental models' respectively. The distinction between a metaphors and a mental model is grey, but for the purpose of this work we refer to metaphors as figures of speech or images and to mental models as representation which include some sort of description of system functioning (see Table 1). Without being exhaustive, cities can be seen as:

• A 'natural' form of human aggregation: ecosystems, organisms, beehives, ant colonies
• An artificial system, antithetic to Nature: machines, infrastructure assemblages, built environments
• An economic entity: engines of prosperity, components (of globalization), competitive economic agents, cathedrals of consumption
• A locus of cultural development: brains, smart city, the place where an intellectual critical mass can flourish
• A locus of artistic development: the 'city beautiful', 'city spectacles'
• A locus of social alienation: rat race/rat maze, a black hole (a sink of resources), a trap (an inescapable monotony of routines devoid of meaningful human interaction).

How can the types of urban metaphors and mental models be used to support urban resilience? As we try to think systematically about this, we relate the urban mental models to the three key concepts in urban resilience (desired function, adaptive capacity and system understanding), extracted from the definition in the introduction, and shown in Table 2.

Table 2: Mapping of models and metaphors against core urban resilience aspects

We acknowledge that the difference between row 2 and row 3 is subtle and perhaps even subjective as we here single out learning processes as being separate from urban systems description rather than an integral part of it. Thus we see them as two varieties of the same type of urban mental model (Type B1 and B2). The reason for this separation is to highlight mechanisms for learning and/or adaptation as particularly important processes in cities.

Type A. Urban mental models that describe an outcome: as above, this is a range of urban metaphors and mental models that can be best described as desired end- goals in urban configurations. We call these 'urban-state mental models' and include, for example, the city beautiful, city spectacles, a locus of social alienation, a black hole, city as a trap, among others. 'Urban-state mental models' describe outcomes of urban processes, decision rules or management actions. As such, they are value laden, because some outcomes can be positive (city beautiful, cultural experience) and other may be negative (social alienation, rat race). We believe that it is important to keep looking for new desired urban configurations. This can be achieved by the urban governance system engaging the public to expand the list of mental models associated with desired functions in Table 2. We also recognize that some of the desired functions support adaptive capacity. For example green, low carbon, water sensitive and walkable cities are likely to reduce pressures on the city. However, all these are partial goals and we need mental models of city functions that embed all the desired aspects of a city, rather than only narrow definitions of urban outcomes. Without this, the city can't be managed as a system, and it is then difficult for urban managers to provide anything but piecemeal solutions. To help urban managers, we need more holistic urban-state mental models of cities. It is also worth noting that it is not always possible to manage a city to delivery outcomes that suit everyone, i.e., there are winners and losers; and so these mental models are sometimes contested.

Type B1. Urban learning and adaptation mechanisms: several of the mental models describe aspects of how a city can adapt and learn, primarily by providing information and data to support decision making. With the exception of the term tactical urbanism, the available mental models to guide the learning processes, i.e., the smart cities, cities as machines etc. are largely focused on the technology and information, and under-represent the social aspect of learning, i.e., social capital. This is not to say that technology is not important for the city, but we argue that technology is insufficient by itself. The city as brains metaphor stands out as it focuses on the creative and cultural aspects of a city (Bettencourt et al., 2007). According to current knowledge, urban change processes are both social and technological in nature (Webb et al., 2017), and so we think urban-learning mental models should be (Bai et al., 2016). Combining the concept of tactical urbanism with the metaphor of a city as a brain, may provide useful results.

Type B2. Describing urban systems: some urban mental models are more explicit at representing a city as a system and how it functions. These include cities as: ecosystems, organisms, beehives, artificial systems, infrastructure assemblages, economic entities and engines of prosperity, among others. These may potentially be turned into computational models. We have also identified urban metaphors that describe the physical processes of the city (urban metabolism), the city in the context of the regional and the global economy (the city as actors), and cities as places where technological innovation occurs (socio-technical systems). The city as a complex adaptive systems (CAS) is a fairly theoretical urban mental model, that provides some guidance in how to explore cities, yet most descriptions of cities as CAS struggle in practice with defining the system, both in terms of boundaries as well as in terms of the required level of detail. Nearly all urban systems mental models are focused on solving specific problems, yet no urban mental model describes the urban system in its entirety.

From a system perspective, these three types of urban mental models differ significantly. The Type B1 and B2 'urban-system mental models' are process oriented and describe complex webs of cause and effect mechanisms. As such, they suggest causal process and thus potential levers for policy intervention. However they are value free, as they say little about what experience they may provide to the citizens. On the other hand, as 'urban-state mental models' can be outcomes of simplistic decision rules, they may be responsive to social processes or governance choices, not first causes. It should also be recognized that the 'urban-state mental models' may be perceived as the outcome of cause and effect relationships that are not always based on solid scientific evidence, so that their validity could be challenged. Furthermore, simple decision rules in complex adaptive systems will tend to have both intended as well as unintended consequences that are often unpredictable. So whilst the Type A type of mental model may be useful, they should be used with caution.

Mental models as building blocks for computational models

Mental models are useful building blocks on which directed urban planning can be supported. However, limitations in human cognition and the immense complexity of urban processes, require computational models in order to provide adequate decision support. See Table 1 for a distinction between mental and computational models. All models reside, implicitly or explicitly, partially or fully, in someone's mind or in the minds of several people, so they are all 'mental' models. A few of them are turned into computational models to expand beyond the limited cognitive functions of humans. This contrasts with a common view of numerical and mental models as an alternative representation of a process. We believe there are a number of benefits in this interpretation (Boschetti, 2015):

1. It demystifies the essence of computer modelling in communicating model results to stakeholders;
2. It demystifies the predictability power of computer models;
3. It clarifies why computer models are not value-free and, like any mental model, necessarily come with underlying worldview, perceptions, and narratives.

Computational models can play different roles to support resilience based assessments and planning, i.e., models can be used to:

1. Develop and communicate a system description by making explicit a set of coherent assumptions about a system's components and processes.
2. Facilitate conversation, and thus develop a shared understanding of issues and solutions by representing divergent views and invite scrutiny and comparison with evidence (Meadows, 1974).
3. Inform what to measure and monitor and how to identify performance metrics or indicators.

To explore the suitability of urban mental models as building blocks for urban computational models to support resilience based urban planning, in Table 2 the relationship between urban metaphors, the embedded worldviews, and their implications for urban resilience are explored.

Table 3: The relationship between selected city mental models and implication for a city's resilience

Reflections: Where to next

This exploration of urban mental models indicates that any mental or computational model of a city is inevitably partial, and that an adaptive governance approach that acknowledges and guides the use of multiple perspectives and multiple mental models is more desirable. This would include surveying the mental models of stakeholders through qualitative research methods. But how can the urban mental models support urban governance?

1. Firstly, we believe that urban-state mental models are useful for guiding monitoring and data collection, because they are often normative and describe a hypothesis about how city configurations are associated with desirable or undesirable consequences. Therefore, these are suitable points of call when choosing an indicator framework.
2. Secondly, to understand the past, current and future challenges of a city requires a combination of data (within decision support frameworks) and deliberation by stakeholders and experts, for example via qualitative research methods like scenario planning. This will enable a first cut understanding of the issues which need addressing, that can reveal and be aided by available mental models.
3. Thirdly, to identify strategies for transforming the city and/or to address particular challenges, it is usually necessary to use a model. The type of model or metaphor to be adopted however will depend on the type of challenge that needs to be addressed.
4. Fourthly, in terms of modelling, current models are patchy, limited and very rarely provide an adequate systems perspective. One way forward would be to create inter-operability of urban models that describe partial issues; i.e. by connecting models that describe health outcomes, traffic congestion, social inequality, urban ecology, urban sprawl, etc. This would indicate that the frontier of dealing with urban complexity is perhaps not in the development of large scale urban models, but by providing the inter-operability of partial urban models.

We also think that participatory modelling (see Table 1) has a considerable potential for describing and understanding a city as a system. This can employ different tools, such as systems dynamics or agent-based representations of a city, that describe the high level properties of an urban system in a stylized manner, albeit without a solid empirical basis. Among these tools, appropriate systems archetypes can be particularly useful. System archetypes (see Table 1) represent the building blocks of dynamical systems (Wolstenholme, 2003). To provide an example, feedback loops are an important concept for resilience because they help understand what reinforces and accelerates or what weakens and impedes change. System dynamics researchers have identified generic structures of feedback loops and their associated rates or delays that can be combined to create narratives of what can or cannot unfold in a system because of dynamical constraints. From Argyris and Senge (Argyris & Senge, 1990) "If reinforcing and balancing feedback and delays are like the nouns and verbs of systems thinking, then the systems archetypes are analogous to basic sentences or simple stories that get retold again and again". Descriptions of ten of these archetypes are presented as a 'family tree' (Kim, 1993). Each archetype is characterized by a generic causal loop diagram, whose dynamic properties are similar across many different kinds of systems and scales. In other words they are intended to be over-arching metaphors suitable for a diverse range of systems. This provides a further typology of metaphors: 'linguistic metaphors', which we have discussed so far and 'dynamical metaphors', as represented by the system archetypes. Technically, a combination of system archetypes interacting with one another is a computer model and can be used to carry out formal, numerical computation. However, an underlying assumption when using system archetypes is that their dynamic properties are robust over a range of possible quantitative parameterizations and thus their qualitative arrangement can have some level of qualitative predictive power. This can be powerful in interactive workshop settings, since archetypes and their results can be communicated via diagrams and analogy, eventually leading to narratives, rather than mathematical formalism.

To summarize, we believe that the next steps in this effort of providing appropriate guidance on urban systems levers and thus informing holistic urban governance, is to:

1. Develop a robust planning framework that embeds a mixed methods approach of iterations between scenarios and modelling capability to support adaptive governance, similar to the Management Strategy Evaluation method (Fulton et al., 2015).
2. In the absence of adequate urban model inter-operability, provide a decision support framework that allows the choice of an appropriate systems archetype to guide participatory modelling activities.
3. To guide the robust planning approach, develop a suitable city indicator framework, based on a review of urban-state metaphors, as described above.

References

Argyris, C. and Senge, P. (1990). The Fifth Discipline: The Art & Practice of the Learning Organization, ISBN 9780385260954.

Bai, X., Surveyer, A., Elmqvist, T., Gatzweiler, F.W., Guneralp, B., Parnell, S., Prieur-Richard, A.H., Shrivastava, P., Siri, J.G., Stafford-Smith, M., Toussaint, J.P, and Webb, R. (2016). "Defining and advancing a systems approach for sustainable cities," Current Opinion in Environmental Sustainability, ISSN 1877-3435, 23: 69-78.

Bettencourt, L.M.A., Lobo, J., Helbing, D., Kuhnert C., and West, G.B. (2007). "Growth, innovation, scaling, and the pace of life in cities," Proceedings of the National Academy of Science, ISSN 1091-6490, 104(17): 7301-7306.

Boschetti, F. (2015). "Models and people: an alternative view of the emergent properties of computational models," Complexity, ISSN 1099-0526.

Fulton, E.A., Boschetti, F., Sporcic, M., Jones, T., Little, L.R., Dambacher, J.M., Gray, R., Scott, R., and Gorton, R. (2015). "A multi-model approach to engaging stakeholder and modellers in complex environmental problems," Environmental Science and Policy, ISSN 1462-9011, 48: 44-56.

Jones, N.A., Ross, H., Lynam, T., Perez, P., and Leitch, A. (2011). "Mental models: An interdisciplinary synthesis of theory and methods," Ecology and Society, ISSN 17083087, 16(1): 46.

Kim, D.H. (1993). Systems Archetypes I: Diagnosing Systemic Issues and Designing High- Leverage Interventions, ISBN 9781883823009.

McPhearson, T., Iwaniec, D.M., and Bai, X. (2016). "Positive visions for guiding urban transformations toward sustainable futures," Current Opinion in Environmental Sustainability, ISSN 1877-3435, 22: 33-40.

Meadows, D.H. (2008). Thinking in Systems: A Primer, ISBN 9781603580557.

Meadows, D.L. (1974). Dynamics of Growth in a Finite World, ISBN 9780262131421.

Meerow, S. and Newell, J.P. (2016). "Urban resilience for whom, what, when, where, and why?" Urban Geography, ISSN 1938-2847, 1-21.

Moglia, M., Cork, S.J., Cook, S., Boschetti, F., Bohensky, E., Muster, T., and Page, D. (2018). "Urban Transformation Stories for the 21st Century: Insights from Strategic Conversations." Global Environmental Change In Press, ISSN 0959-3780

Webb, R., Bai, X., Smith, M.S., Costanza, R., Griggs, D., Moglia, M., Neuman, M., Newman, P, Newton, P, Norman, B., Ryan, C., Schandl, H., Steffen, W., Tapper N., and Thomson, G. (2018). "Sustainable urban systems: Co-design and framing for transformation." Ambio, ISSN 0044-7447, 47(1): 57-77.

Webb, R., Bai, X., Smith, M.S., Costanza, R., Griggs, D., Moglia, M., Neuman, M., Newman, P., Newton, P., Norman, B., Ryan, C., Schandl, H., Steffen, W., Tapper, N., and Thomson, G. (2017). "Sustainable urban systems: Co-design and framing for transformation," Ambio, ISSN 0044-7447, 47(1): 57-77.

Wolstenholme, E.F. (2003). "Towards the definition and use of a core set of archetypal structures in system dynamics," System Dynamics Review, ISSN 1099-1727, 19: 7-26.