National Academies Press: OpenBook
« Previous: Summary
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×

1

Why Models?

We live in a world of numerous, varied, and evolving threats to national security, including terrorism, proliferation of weapons of mass destruction, scarcity and disruption of food and water supplies, extreme weather events, natural disasters, and regional conflicts and disputes (Clapper, 2015; NIC, 2012). Many of these threats have an important spatial component, and understanding where exactly something is, what it is, and why it is important is the job of the National Geospatial-Intelligence Agency (NGA). NGA analysts evaluate imagery and remote sensing and other data on the land, ocean, and atmosphere to create geospatial intelligence in support of national security, disaster response, and humanitarian assistance (see Figure 1.1).

Image
FIGURE 1.1 An NGA analyst at work. Analysts acquire, process, and analyze imagery and other geospatial information from a variety of classified and open sources and deliver information products and services to policy and decision makers. SOURCE: http://gcn.com/articles/2014/11/20/nga-map-of-the-world.aspx.
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×

NGA grew out of a merger of several defense and intelligence organizations that provided mapping, charting, imagery analysis, and geospatial information services.1 Initially called the National Imagery and Mapping Agency, the name was changed to NGA in 2004 to mark the emergence of the discipline of geospatial intelligence. The term geospatial intelligence is defined as “the exploitation and analysis of imagery and geospatial information to describe, assess, and visually depict physical features and geographically referenced activities on the Earth.”2 Geospatial information includes collections of data items, each referenced to a location on the surface of Earth. It may describe places at multiple scales (e.g., countries, provinces, counties, and neighborhoods); physical objects (e.g., buildings and vehicles); and spatiotemporal activities (e.g., movement of people), relationships (e.g., networks among people, institutions, and places), and patterns (e.g., hotspots and co-occurrences).3 These attributes are relevant to both mapping and modeling.

NGA has developed some models that characterize geophysical features and phenomena (e.g., land and seafloor terrain, and magnetic and gravity fields [see Figure 1.2]) and, increasingly, the human geography (e.g., distribution, alliances, and hostilities among ethnic groups across a country). However, NGA envisions developing a broader modeling and analysis capability that will enable analysts to search for spatial and temporal patterns; make inferences from those patterns; predict future political, economic, and military threats to the United States; and evaluate options and consequences around the world. This added capability would be used to better inform policy choices, action plans, and contingency strategies that depend on geospatial intelligence.

Image
FIGURE 1.2 Global map of free air gravity anomalies (difference between measured and reference gravity values, corrected for elevation, in milligals [mGal]), computed from NGA’s Earth Gravitational Model 2008. Features with relatively high mass (e.g., mountains and subduction zones) have high measured gravity values relative to the reference value, and a positive gravity anomaly (warm colors). NGA uses these data to support navigation systems, mapping, and surveys. SOURCE: NGA, http://earth-info.nga.mil/GandG/wgs84/gravitymod/egm2008/anomalies_dov.html.

___________________

1 NGA Reference Chronology, available at https://www.nga.mil/About/History/Pages/Educational-Resources.aspx.

2 10 USC § 467.

3 Richard M. Medina and George F. Hepner, A note on the state of geography and geospatial intelligence, https://www.nga.mil/MediaRoom/News/Pages/StateofGeographyandGEOINT.aspx.

Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×

At the request of NGA, the National Academies of Sciences, Engineering, and Medicine established a committee to describe mathematical, numerical, and statistical models and spatiotemporal analytical methods used to understand real-world systems; to determine what would be required to make these models and methods useful for geospatial intelligence; and to identify supporting research and development for NGA (see Box 1.1). NGA offered two pieces of guidance for addressing these tasks. First, a wide variety of public-domain, proprietary, and classified models and methods are likely relevant, and NGA asked the committee to focus on those in the public domain, which have largely been developed by university researchers. Second, the study was unclassified in its entirety, and little information was available on NGA’s current capabilities and future needs. NGA provided an overview of the agency and geospatial intelligence, as well as two example intelligence scenarios (see Box 1.2). The committee used this information, national security threats with a geospatial component (e.g., Clapper, 2015; NIC, 2012), and members’ experience with NGA and other defense agencies to guide its selection of potentially relevant models and methods, and its suggestions for NGA research and development needs.

DEFINING MODELS

Generally, a model is a simplified representation of a real-world system that enables investigation of system properties and behaviors. In this report, a “model” means a mathematical or numerical model that can be run on a computer. It typically requires inputs that specify initial conditions, boundary conditions, and/or model parameters to produce an output. A model could be as simple as a linear mapping from input to output (xax + b), or as complex as a climate model that includes multiple processes operating at multiple and temporal and spatial scales, evolves the state of Earth’s ocean and atmosphere over centuries, and runs on a supercomputer. Such models are used to extract explainable insights about the system, to enable prediction of future outcomes, or to aid in decision making by simulating multiple “what-if” scenarios for consideration.

Models tend to fall somewhere on the spectrum ranging from data driven to theory driven. Data-driven empirical models (e.g., regression and classification models from the fields of statistics and machine learning) are designed to ingest data and efficiently estimate model parameters, capturing dependencies and correlations between system features. Theory-based process models encode causal connections between states, entities, or subsystems, describing system evolution and/or response to changing conditions. The appropriate level of empiricism depends

Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×

on a number of factors, such as the availability of system observations, the computational demands of the model, the maturity of the theory, and the needs of the model-based investigation.

MODELS FOR UNDERSTANDING REAL-WORLD SYSTEMS

The notion of building models of real-world systems has a long history. Among the first predictable, recurring phenomena humans tried to model were astronomical in nature, including the diurnal cycle, the passing of the seasons, and the movement of planets in the night sky. Early astronomers, including Aryabhata (born in India in 476) and Tycho Brahe (born in Denmark in 1546), carried out the immensely complex computations by hand. The advent of digital computing in the mid-20th century offered a way to perform massive computations based on mathematical formulas, and revolutionized modeling capabilities. Weather and climate modeling is a classic example of how digital computing transformed an entire field (see Box 1.3).

The current situation in social system modeling is not unlike weather forecasting 35 years ago, when computational methods first began surpassing expert judgement (Edwards, 2010). For example, the Federal Reserve Board now relies on both experts (board members) and models (e.g., FRB/US) to estimate the near-term prospects for the U.S. economy, which are then used to decide on policy. The performance of both the experts and the models has been checkered. In the 2007–2009 financial crisis, the FRB/US model correctly predicted the steep drop in housing prices, but substantially underestimated the rise in unemployment. Expert predictions of unemployment were also too optimistic, even while the crisis was unfolding. A variety of computationally sophisticated social system modeling efforts are now under way in economics (e.g., Delli Gatti et al., 2011), finance (e.g., Geanakoplos et al., 2012), and other policy realms (e.g., state stability and epidemics). Consequently, prospects for these domains to follow the trajectory of weather model development are strong.

THE MODEL-BASED INVESTIGATION PROCESS

An investigation of a real-world system is typically an iterative process, framed by the central tasks of (1) identifying key questions to be addressed, (2) scoping the investigation, (3) exploiting models to make inferences about the key questions, (4) assessing the model-based analyses, and (5) revising any of these steps as necessary.

Identifying Key Questions or Features to Explore

A model investigation is focused on particular questions or features of a system. These questions drive how models will be exploited, ensuring that the investigation yields information relevant to the users. Different user requirements may lead to fundamentally different models of the same phenomenon. Consequently, it is important to choose the questions carefully. For example, the broad intelligence question in NGA’s megacities scenario (How will worldwide urbanization trends affect regional political, economic, and security environments? [see Box 1.2]) likely encompasses several model-based investigations on different aspects of the issue (e.g., the probability and impact of heat waves and the impact of changing demographics on crime).

A related task is to ask what value modeling will add in addressing the key questions. In some cases, the complexity of the system relative to the sophistication and fidelity of available models, or the paucity of relevant observations of the system, makes the effort nearly impossible. Choosing investigations where quantitative modeling is likely to shed light on the key questions is crucial.

Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×

Scoping and Planning the Investigation

The real-world systems associated with NGA’s example intelligence scenarios contain a large number of heterogeneous subsystems (e.g., geophysical, environmental, cultural, and economic), and the models used to investigate them are computationally challenging, are data rich, and require an arsenal of analytic approaches. Scoping includes determining what models, data, analytical methods, subject-matter expertise, computational resources, and other resources are required for the investigation. For example, different models have different capabilities, such as their ability to capture particular aspects of reality, and these capabilities must be compared with the characteristics needed from the analysis to decide on a modeling approach. Scoping is intimately connected to plans for combining these ingredients to provide useful information about the key questions. The speed and accuracy of required results, the availability and maturity of existing models, and the availability and completeness of data also affect scoping.

Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×

Exploiting Models

A model-based investigation may involve developing a new model, setting up and running an existing model, combining or coupling existing models, and/or analyzing results from previous model runs or analyses. The models or model products are then incorporated into an analysis, typically involving data and computation, to connect the model to the real-world system under investigation and to produce insights, predictions (and uncertainties), or plausible outcomes for some features of the system. For NGA investigations, the analysis will be influenced by the spatial, space-time, or network structure that is present in geospatial data. Key characteristics of such data include autocorrelation (e.g., the properties of nearby locations tend to be similar) and spatial heterogeneity (i.e., the phenomenon being modeled varies with location).

Model exploitation is at the heart of any investigation, and it is explored more broadly in Chapter 3. It is worth pointing out here that the strengths and limitations of the models used are inherited by the larger investigation.

Assessing the Credibility of Model-Based Insights

A model differs from the real-world system it seeks to represent for several reasons, including omitted or inadequate representation of system processes, errors and uncertainties in the data used as input to the models, and coding errors in the models. Any model-based investigation must assess the impact of these uncertainties on our inferences about the real-world system and communicate this uncertainty to decision makers. Ideally, this assessment takes place throughout the development of the model, addressing the question of whether the behavior of a model sufficiently matches the behavior of the real-world system to address the key question(s) of the investigation. With mature models and sufficient data, the assessment process often covers the adequacy of the model’s representation to the real-world system (validation) and the quantification of model uncertainty. For models with exploratory and extrapolative elements or insufficient data for comparison, assessment processes may be more qualitative, particularly in situations where uncertainty is large and difficult to quantify. Evaluation processes might also involve comparing predictions or outcomes from different models, especially if they use different conceptual modeling approaches.

Making Revisions

Information and insight gained throughout the investigation is used to revise any of the elements mentioned above—including the formulation of key questions, the application of resources, the computing infrastructure used, and the data collected—to improve the model and the quality of analysis results. For example, the empirical elements of a model may be upgraded as new data become available. Thus, the model representation itself may evolve in response to new data.

STRENGTHS AND LIMITATIONS OF MODELS

Models simplify relationships and omit or aggregate some features and processes of the real-world systems they are representing, depending on the key questions of the investigation. This abstraction is precisely what makes models immensely useful, because it reduces the number of processes that may be acting or changing and thus provides clearer insight on the most important aspects of the system. In addition, a model can be explored in ways the real-world system cannot. For example, sensitivity studies can be carried out to determine how varying key model inputs causes model outputs to change. Simulation-based experiments can be carried out to assess the model’s response to particular input forcings or management strategies.

Models have led to increased understanding and produced accurate predictions in a wide variety of physical

Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×

systems, including weather (Simmons and Hollingsworth, 2002; see Box 1.3), nuclear physics (Hendricks et al., 2008), and astrophysics (Fryxell et al., 2000). Even a simple, empirical model can lend insight on a complicated system, as illustrated by Francis Galton’s use of correlation and regression to understand features of heritability in the late 1800s (Stigler, 1986). On the other hand, models did not anticipate the accumulation of local influences and external forcings that caused the collapse of the Grand Banks fishery in 1992 (McGuire, 1997) or the financial crisis in 2008 (Taylor, 2009).

Models also have limitations in capturing important behaviors of complex systems. Many of the systems of interest to NGA have a large number of interrelated and autonomous subsystems, and both these subsystems and the larger system can adapt. In such complex adaptive systems, complexity can be thought of as the potential for emergent behavior to appear and for small changes to have unforeseen and potentially enormous consequences (Holland, 1992; Lansing, 2003). Such systems often exhibit nonlinear behavior, and the same action at different points in the system’s history may have different results.

The gap between model and reality can be substantial when the real-world system itself is evolving toward a new regime that is unlike previous experience. For models that have been tuned to perform well in comparison to past data, there is a danger that the model and reality will diverge in a new regime, even if the model has adequately tracked the system in the past. Model-based predictions of climate change embody this concept. Many climate models can accurately simulate the global temperature record, but their predictions of warming for the 21st century can differ by a factor of 2 or 3, depending on how the models handle aerosols and greenhouse gases, which contribute to warming in different ways (Kiehl, 2007). Hence, accuracy in reproducing past conditions is no guarantee of accuracy in projecting future conditions, since the combined (and potentially nonlinear) effects of two different forcings may not be fully realized in the historical data alone. A change in regime also affects data-driven models, which are trained to learn correlations and patterns from existing data.

In addition, prediction accuracy declines over longer timescales because of complex or chaotic system behavior. The limits of predictability in fully specified, completely deterministic systems have been demonstrated by Lorenz (1963). Chaotic systems might also contain multiple regimes and exhibit transitions (abrupt changes from one equilibrium to another), tipping points (changes from one stable regime to another), hysteresis, and path-dependent behaviors. Figure 1.3 shows a trajectory in phase space which appears to orbit around one center of attraction, but a very small displacement can shift the trajectory to an orbit around an entirely different center of attraction. A small error in the initial condition can thus shift the system into a different equilibrium. Although such systems pose challenges for prediction and understanding, they also exhibit tendencies and behaviors that can be explored with models.

Model initial conditions, parameters, or forcings can be perturbed to develop a range of potential trajectories, which can be helpful for exploring possible future outcomes, testing analysis approaches, decision making, or training. For example, potential trajectories are often combined with the perceived cost or benefit of each trajectory and the user’s attitude about decision making under uncertainty (e.g., Howard, 1968; Keeney, 1982; Raiffa, 1968). Model-based trajectories have been used to guide decision making in a variety of application areas, including climate mitigation (e.g., Drouet et al., 2015), smallpox interventions (e.g., Ferguson et al., 2003), and terrorism response (e.g., Rosoff and von Winterfeldt, 2007).

Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Image
FIGURE 1.3 The Lorenz butterfly, emblematic image of chaos. The figure depicts the strange attractor that can unpredictably and abruptly switch between multiple regimes. SOURCE: Wikimedia.

Because most decision makers or other users have not been involved in the modeling, it is particularly important to communicate the strengths, weaknesses, and overall level of uncertainty of the model-based investigation. As cautioned by Chesshire and Surrey (1975):

Because of the mathematical power of the computer, the predictions of computer models tend to become imbued with a spurious accuracy transcending the assumptions on which they are based. Even if the modeler himself is aware of the limitations of the model and does not have a messianic faith in its predictions, the layman and the policymaker are usually incapable of challenging the computer predictions. Depending upon the modeler’s credibility, a dangerous situation can arise in which computation becomes a substitute for understanding about problems and relationships, and a substitute rather than an aid for policy choices.

Even the results from simple models are subject to misinterpretation if their many assumptions and uncertainties are not conveyed clearly. Communicating uncertainty of scientific results remains an active area of research (e.g., Fischoff and Davis, 2014; Morgan and Henrion, 1990; NRC, 2006, 2007a).

ORGANIZATION OF THE REPORT

This report describes models and analytical methods that are potentially relevant to NGA and discusses how they could be developed, used, or adapted for geospatial intelligence purposes. Chapter 2 illustrates the power of models and methods to tackle issues of potential interest to NGA. Chapter 3 describes the components of a model-based investigation, including the models themselves, the data linking the model to the real-world system,

Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×

methods for analysis and model assessment, and the computational infrastructure. Example questions are posed to illustrate how NGA could think about these components. Chapter 4 covers models and methods likely to be of particular interest to NGA, the state of the art in these models and methods, ways to make them useful to NGA, and research and development that could help NGA adapt, use, and maintain the models for geospatial intelligence purposes. Additional detail on methods for combining models and on computation appears in Appendixes A and B, respectively. Biographical sketches of committee members are given in Appendix C, and acronyms and abbreviations appear in Appendix D.

Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×

This page intentionally left blank.

Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 9
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 10
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 11
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 12
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 13
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 14
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 15
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 16
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 17
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 18
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 19
Suggested Citation:"1 Why Models?." National Academies of Sciences, Engineering, and Medicine. 2016. From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities. Washington, DC: The National Academies Press. doi: 10.17226/23650.
×
Page 20
Next: 2 Illustrative Models »
From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities Get This Book
×
Buy Ebook | $34.99
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

The United States faces numerous, varied, and evolving threats to national security, including terrorism, scarcity and disruption of food and water supplies, extreme weather events, and regional conflicts around the world. Effectively managing these threats requires intelligence that not only assesses what is happening now, but that also anticipates potential future threats. The National Geospatial-Intelligence Agency (NGA) is responsible for providing geospatial intelligence on other countries—assessing where exactly something is, what it is, and why it is important—in support of national security, disaster response, and humanitarian assistance. NGA’s approach today relies heavily on imagery analysis and mapping, which provide an assessment of current and past conditions. However, augmenting that approach with a strong modeling capability would enable NGA to also anticipate and explore future outcomes.

A model is a simplified representation of a real-world system that is used to extract explainable insights about the system, predict future outcomes, or explore what might happen under plausible what-if scenarios. Such models use data and/or theory to specify inputs (e.g., initial conditions, boundary conditions, and model parameters) to produce an output.

From Maps to Models: Augmenting the Nation's Geospatial Intelligence Capabilities describes the types of models and analytical methods used to understand real-world systems, discusses what would be required to make these models and methods useful for geospatial intelligence, and identifies supporting research and development for NGA. This report provides examples of models that have been used to help answer the sorts of questions NGA might ask, describes how to go about a model-based investigation, and discusses models and methods that are relevant to NGA’s mission.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    Switch between the Original Pages, where you can read the report as it appeared in print, and Text Pages for the web version, where you can highlight and search the text.

    « Back Next »
  6. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  7. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  8. ×

    View our suggested citation for this chapter.

    « Back Next »
  9. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!