and again in this field guide, is a key task in controversy mapping.
Links may show alignments. They may reveal the company an actor keeps, or unlikely bedfellows. Particularly when displayed through the techniques described in the sections on visual network analysis, the investigation of links could indicate actor standing. There are hierarchies in networks, and links furnish certain actors with prominence and centrality. “Orphan” websites, without inlinks, did not make it onto the maps, however, so non-association would require another technique (such as mentions mapping, discussed below).
More broadly, these webs of actor links, in other words, were considered networks. Linking actively brings networks into being. They create a space of interaction and circulation, the “magmatic” space of alliances and oppositions described in the book. They each may be characterized by the material they circulate and the activity they display. Rather than structural, such as “old boy networks,” we found them to be dynamic, and continually in need of glue. These networks, we thought, were held together by what they principally circulate. Debate networks circulate positions, scandal networks accusations, issue networks urgency, summit networks announcements, and protest networks mobilization.
The rates at which these networks circulate the material, through the updating of their websites, could be considered a measure of dedication or care. The “health” or hardiness of the issue, protest, and so forth could be monitored. In order to do so, the maps went live.
We would build issue tickers and other displays to show an issue’s current state. In one example, we monitored NGO campaigning behavior, and when NGOs would leave an issue, in the sense of ceasing its advocacy, the issue’s state would decline. One could similarly monitor a controversy, and when it is the hottest, or in full motion, that is the occasion to map and chart it.
Second, we mapped mentions. In order to capture an actor’s partisanship or issue commitment, we learn from how actors mention each other as well as each other’s issue language. It is an issue mapping exercise, where, for example, we asked, which actors in the climate change space mention the skeptics? Here the idea is to gain a rough indication of skeptic-friendly sources. (Sources skeptical of the skeptics also appear, so close reading and annotation are essential.)
Querying search engines, and interpreting actor mentions, are not particularly slow methods, but they are not single-click operations either. Their exploitation demands a reflexive attitude, including the many methodological considerations laid out in the chapter on the collection and sorting of digital records. Research browsers are installed, settings are changed, and multiple queries are made as one reconsiders the use of search engines, transforming them from a consumer information appliance into a research device.
We considered the placement of the mentions. This is the study of what is referred to as “source distance.” How far from the top are the mentions? It is also a consideration of the significance of the actors who can be seen to sympathize or align with them.
In order to produce a coarse measure of partisanship among significant actors in an issue space, the search engine is queried, but in two steps, first, for the issue, “climate change.” The second set of queries is made for the names of skeptics or skeptical organizations in each source returned in step one. This is the point where the search engine has been repurposed for research.
We note a source’s affinity toward an actor, but we also would like to draw comparisons between the actors. Who appears to be the most sympathetic? There is a word-clouding technique, where one keeps the order of the sources returned and resizes the actor names according to the number of mentions of the other. One can have clouds (which also could be called maps) that show the results of an analysis of a single actor or a constellation of many, for example, which actors mention a set of skeptics. One can also show the most significant skeptic.
Networks also could be said to have an agenda or a set of commitments. Certain issues may be suffering, others thriving. If one demarcates a network through hyperlink analysis, relying on an authoritative list or surveying actors or funders, one can harvest each actor’s online issue list, merge the lists, and query every actor for every issue (in our repurposed search engine). The resulting “issue cloud,” whether in alphabetical order or arrayed by mention frequency, provides both the network’s overall agenda as well as the distribution of commitment. Are children’s rights doing well within the global human rights issue network? Who’s particularly dedicated to it, and who is much less so? Claimed commitments also may be tested. An organization may have a beefy issue list but pay attention to a few over the others listed.
Third, the maps have effects. As this field guide makes clear from the outset, to be a mapmaker is to assume some form of complicity. Link maps could make actors desirous of others’ sets of inlinks. They may be envious of another actor’s positioning either overall or within a particular cluster. Having viewed the maps, actors might begin to solicit links. They could consider employing the services of a link farm, just as celebrities and others purchase likes or followers to enhance their social media status. Sudden movement around the map could capture actor link optimization, or could be a sign of improved showing, however gained. Mapping becomes temporal, with a series of scheduled snapshots.
But the maps are themselves part of the issue space, certainly when actors incorporate them into their PowerPoint presentations. Just like the geographical maps discussed in the chapter on the representation of controversies, debate maps are not neutral descriptions of a territory, but instruments in the hands of social actors. In one project, working in a data sprint similar to the ones described in this book, we collaborated with FairPhone, the project originated at the Waag Society, Amsterdam, to create a “more sustainable smartphone.” They had a map, or diagram, that shows the stages of life of a smartphone that emphasizes the mining and recovery of minerals. In their presentations they raise awareness of the mineral trade, also discussing “conflict minerals.” They argue that there is scant information available about those raw materials in the various phases of a smartphone’s life from production, sale, and use to recycling. With mineral mention mapping, we annotated their map by adding an awareness layer (see figure 0). It could be inserted into the PowerPoint presentation, after the original stages of life diagram. It empirically hardens the point made by FairPhone, but also indicates some signs of awareness, perhaps from their own dedicated efforts (though we did not test that supposition).
Finally, there are a number of precepts that the mapping of issues and networks with digital methods have followed that are decidedly Latourian (apart from having precepts in the first place): multiply the maps, instead of seeking to make the mother map and have it serve as the endpoint. Multiple map-making is the research that precedes and enables making findings or developing a narrative of the results. Make the maps reversible, so one can follow content or data points back to their origins. These key precepts of “second-degree objectivity” and “datascape navigation” (described in the chapter on representation) were imported from ANT into digital methods. A recent digital methods tool, 4CAT, for collecting and analyzing 4Chan, Reddit, TikTok, Telegram, and Instagram data, not only queries the social media platforms (or archives of them), but also saves each analytical sub-operation per map. In that sense it allows for the deconstruction of the map. The last precept to be mentioned is the retention and study of actor issue language. While the respect for actor accounts is anthropological in its origins (as argued in the chapter about actor-networks), the issue language also serves as queries. A preferred technique, among the many introduced over the years on collecting and sorting digital records, is to compare the resonance (together with its publics) of competing issue languages, also known as programs and antiprograms, which are like campaigns but without the outward campaigning.
While digital methods have learned from ANT and controversy mapping, this book represents a deepening and maturing of the relationship. It also demonstrates the refinement of many of the techniques that were at the heart of the first wave of web research, described above, when issue crawlers and Google scrapers were routinely deployed for mapping. Welcome to the new wave.
Richard
