Abstract. In analysing controversies and debates—which would include reviewing a literature in order to plan research, or assessing intelligence to formulate policy—there is no one worldview which can be mapped, for instance as a single, coherent concept map. The cartographic challenge is to show which facts are agreed and contested, and the different kinds of narrative links that use facts as evidence to define the nature of the problem, what to do about it, and why. We will use the debate around the invasion of Iraq to demonstrate the methodology of using a knowledge mapping tool to extract key ideas from source materials, in order to classify and connect them within and across a set of perspectives of interest to the analyst. We reflect on the value that this approach adds, and how it relates to other argument mapping approaches. p. 2

Hypertext maps from this analysis: www.kmi.open.ac.uk/projects/compendium/iraq p. 2

The specific hypothesis we set out to explore in this case study was that knowledge mapping tools could help as an analyst’s tool for making sense of published contributions to the Iraq debate:

• for a given source article: mapping tools should help to clarify (at some level of granularity, dependent on the analyst) the contributions it claims to make and its argumentative structure
• for the ‘gestalt’ of the whole corpus: mapping tools should help to clarify the cross-connections and emerging themes which one would expect someone with a grasp of the debate (as expressed in the articles) to have, and communicate clearly. p. 3

We therefore introduce and reflect on:

• the product: a set of hypertextually linked knowledge maps of the Iraq debate, accessible via a specialist hypermedia tool, and via the Web
• the methodology: how this artifact was constructed
• analytical support: how well the tools assisted the analyst
• reading support: how well the tools assist the reader p. 3

The initial reference for the analysis was the paper “One war, many theories” by Michael Cohen (2005). He reviews the fundamental positions of pro-war and anti- war commentators, and distills from these some themes and questions. Cohen asks “How can we do justice to the multiplicity of positions on the war?”, and proposes three concepts to organise the body of arguments:

• Power, defined as the capacity to produced intended effects
• Degree of institutionalisation, or the degree to which certain values and procedures stemming from them are embodied in a regulatory environment (impacting the role of organizations such as the UN)
• Legitimacy, the moral virtues of a certain act or value such that it finds affinities across a broadly defined populace or societal grouping p. 4

As detailed below, we focused on two issues as a mini-template to organize the ideas:

• What were the causes of the Iraq invasion?
• What are the consequences of the war? p. 4

This case study was conducted as part of GlobalArgument.net, a project we initiated in 2005 as a vehicle for systematically comparing computer-supported argumentation tools through argumentation experiments: participants agree on a topic for debate, a set of source documents from which everyone will work, and a schedule for modelling, publishing and analysing the outputs. We are grateful to Peter Baldwin, co-founder of GlobalArgument.net, and Michael Cohen for collating these articles. http://kmi.open.ac.uk/projects/GlobalArgument.net p. 4

As with any cartographic project, we were aiming to create a consistent visual language. Moreover, since we were creating interactive, hypermedia maps, we also needed to create a set of interaction design conventions (Figure 1). These evolved through the analysis, and were summarized in the opening map to assist the reader. p. 5

Figure 1: Explanation of how to read and navigate nodes in maps. The map icon (blue) show comments(*), number of connections to other maps (4) and total of nodes in the map (16). The node map (pink) show tags and number of connections to other maps (4). We started by defining a top level node tagging scheme based on (i) Cohen’s framework of Power, Institutions, and Normsm abd (ii) our Issue-template focused on Causes and Consequences of the war. Over the course of the exercise, as in any qualitative data analysis process, the tag-based coding scheme evolved as we engaged with the material, classifying and reclassifying it until the tag scheme was applied consistently (Table 1). p. 6

An article map for each of the 25 documents was constructed. Text fragments were dragged and dropped frrom the article into Compendium, classifying, linking and tagging each node (Figure 2). p. 7

Isenmann and Reuter (1997) describe 5 steps to structure arguments using IBIS: 1. Identifying issues, positions and arguments 2. Activating external knowledge sources, select data, statistics, concepts, 3. Creating relations 4. Navigating through the knowledge network 5. Reorganizing the issues network p. 8

However, these steps are not linear (e.g. relations may be made before sourcing related data. Moreover, in documents (as in speech), not all of these elements are either explicit, or occur in that order. Authors do not always start with focused questions. They may start with the main proposition, concept or data; and questions can arise during the document. It is the analyst’s task to convert the prose into a map that shows the core issue(s), possible responses to them, and argumentation for and against them, drawing on data. We discuss later the variable levels of reconstruction that the analyst may bring to this mapping. p. 8

Next, we create gestalt maps to show connections across article maps around themes of interest: causes and effects of the war, and around Cohen’s organising themes. For instance, in order to create a map of Pro-War proponents on the theme of Power, we filter the database using Compendium’s search tool to extract nodes tagged with Pro-War and the three types of Power tag (Figure 4). p. 9

Finally, we organised gestalt maps around the question How could the Iraq invasion be understood?, in which we use issues around the war’s causes and effects, and Cohen’s Norms (ethics), Institutions and Power configurations (Figure 6). p. 11

In our view these knowledge maps have valuable notational properties (the visual language, whether on screen or paper) combined with interactive properties (the particularities of mapping within a specific software environment), a distinction made in various ways by Green (1989), Suthers (Chapter X) and many other diagrammatic reasoning researchers. We would highlight the following distinctive attributes for analysts and readers:

• From text string to visualised, database object. When we extract key sentences from articles, we collate them not merely as text strings (e.g. in a wordprocessor) but convert them into addressable nodes that can be spatially positioned, assigned an icon, linked, tagged, have other nodes placed inside them (if we make them a Map or List container node), and tracked by the system as they are pasted into multiple views. This is similar to qualitative data analysis tools for transcript analysis, but via a much stronger visual interaction paradigm.
• From implicit to explicit structure. As argued by many other proponents of visual modelling and argument mapping, there is value in making explicit and inspectable previously implicit structure in a piece of prose, if meaningful patterns can be perceived directly. One can immediately see the presence of different Issues, Positions and Arguments for/against, the presence of tags, the ‘weight’ of a map (how many nodes inside it), and the level of node transclusion. The power of visual patterns increases with the systematicity of the map layout, which derives from greater formality in the modelling process—a theme to which we return below. Although we started from Cohen’s principles, the mapping’s contribution to grasping the gestalt of the debate rests on how we model connections between individual maps of articles. We are making an interpretive move that goes beyond Cohen’s analysis when we extract a quote, and classify, transclude, tag or link it as a node, since this changes the shape of the digital space along one or more dimensions.
• Multiple perspectives. The new finer granularity of chunking ideas as nodes, combined with tagging of important facets, makes possible the easy extraction of different node clusters for the creation of gestalt maps that convey different dimensions to the controversy. p. 14

As with any structured modelling methodology, the point of investing this effort is to add rigour to the analysis. However, there is a cost/benefit tradeoff: mastering this intellectual discipline is a new literacy that takes effort—literally, “Lots of Argument Mapping Practice” (Chapter X). In our view, the knowledge mapping of the Iraq Debate, whilst still requiring intellectual discipline and close reading, required less cognitive effort than detailed Araucaria/Rationale style argument analysis, to effect construction of a network with some valuable affordances. p. 15

Dialogue Mapping template. Conklin (2006) identifies seven issue types that we were using implicitly throughout the analysis in both article and gestalt maps, but which could be used more consciously and systematically to ensure balanced coverage of the whole debate (Figure 8) p. 16

Expose the argumentation substructure. We can build on the work of our argument mapping colleagues, as introduced above, by integrating aspects of their visual languages into the Issue-centric deliberation scheme at the heart of our approach. As shown in IBIS, we can link two nodes with a supports or challenges link, but this does not illuminate the sub-structure of the argumentation. What kind of argument is being made, and is this fallacious, or incomplete? When mapping another controversy, we have shown how supports/challenges links in a Rationale-like argument tree in Compendium, can be further expanded to show the argumentation scheme (Figures 9 and 10). p. 17

This chapter has focused somewhat on the rational, critical analysis of information and argument connections (see also Ohl’s Chapter X). However, the approach we are developing emphasises a simple visual language that can be used effectively in real time to capture and reflect back a wide variety of deliberative moves, with its roots in facilitating dialogue that is owned by all stakeholders (Conklin, 2006; Selvin, et al., 2002; Papadopoulos, 2004 Selvin, Chapter X). The vision of our ongoing Hypermedia Discourse research programme 5 is to create knowledge cartography tools and practices that integrate heart and mind. We need both critical thinking and open listening as we strive collectively to make sense of, and act on, the complexities and controversies now facing us. p. 20

8. References p. 20