Methodology

Developing a method is critical to the success of a digital project.  For “Power of Attorney,” this required the interdisciplinary collaboration of humanists, geographers, and data specialists. The first step in the “Power of Attorney” project was to copy or digitize approximately one thousand letters of attorney from the districts of Teposcolula and Villa Alta, Oaxaca – two of Oaxaca’s largest multilingual overwhelmingly indigenous regions — for the period 1700-1852 in the Archivo Histórico Judicial de Oaxaca (AHJO) and the Archivo Histórico de Notarías de Oaxaca (AHNO).  Our pilot for the project focuses on letters of attorney from the AHJO for eighteenth century Villa Alta.  The other visualizations remain to be done.

The next step was to decide how to organize the data contained in the letters of attorney.  It is impossible to overstate how important the organization of data is to the success of a digital project: accurate, standardized data provides the basis for any map or visualization.  Producing data in this form is always more complicated and time consuming than one anticipates at the outset.

After reading dozens of letters of attorney for the kinds of information that they contained, we decided to extract the maximum data points possible, and organize them into categories in an Excel spread sheet (Table 1).  Since we wanted to tell a story that unfolded in space in order to highlight the engagement of remote native villages with faraway imperial institutions, we decided to create three overarching categories: people, places, and notarial activity. Within the category of people, we divided the data into indigenous grantors of power or attorney and grantees (legal agents).  Beneath the umbrella category of grantors, we recorded whether they were cabildo officers granting power of attorney on behalf of a pueblo, or individuals like a cacique, or wider groups, such as multiple cabildos granting power of attorney on behalf of a coalition of pueblos.  We also made note of any markers regarding the social status of the individual grantors and the municipal offices that they held.

With regard to the category of grantee, we recorded whether the legal agent was a local indigenous “legal entrepreneur,” a cacique, parish priest, or titled lawyer in Mexico City or Madrid.  Some of these men were untitled paraprofessionals (this was true of the indigenous legal agents), and others had official titles such as “abogado de la Real Audiencia,” “procurador de número” or “agente de negocios.”  These titles when provided are noted in the excel sheet. We also noted whether indigenous grantors gave power of attorney to multiple agents in different places, presumably to maximize access to different courts in different locales.  For both grantors and grantees, we recorded places of origin or residence.

With regard to notarial activity, we recorded whether the letters of attorney were for generalized litigation (a “poder general”) or a specific case or legal purpose (a “poder especial”).  If a poder especial, we noted the nature of the litigation.  Whenever a cabildo granted power of attorney on behalf of a community, we counted it as a single act of power of attorney since the native authorities who signed the letter of attorney represented one legal entity (their corporate community).  Whenever a letter of attorney was authorized by more than one grantor – the cabildo officers (native municipal officials) of five communities, for example — we counted the letter as five individual acts of power of attorney, thereby giving proportionate weight to the notarial “event.” We organized all of this data chronologically, based on the date on which the letter of attorney was notarized.  Later, we added additional data about the villages in the Villa Alta region, including ethno-linguistic identity, population, and administrative status.  We took this data from secondary literature on the region.[1]

Once the excel sheet was done, we standardized the data, and decided which data categories were going to be expressed in the maps and visualizations.  We created many more categories of data than we could express on the maps and visualizations.  In this regard, data visualization – like a written narrative – is a process of selection and exclusion. We have included the entire excel sheet for 18^th century Villa Alta so that students or scholars who are interested in pursuing further research may do so (see the tab for Project Data).

In order to make the maps, which are at the heart of the project, we used Geonames to locate the coordinates for the places listed in the excel sheet so that they could be plotted on a map.  This was more difficult than we had initially anticipated given that some places disappeared over time, the orthography or place names changed over time, as did administrative divisions.  Eventually, we were able to find coordinates for all places, give them unique place identifiers, and create our own gazetteer for the region.  We also gave unique identifiers to grantees.  For the case of grantors, most of them were cabildos (municipal councils) granting power of attorney on behalf of their communities, which meant that anywhere from 8-9 names were listed as signatories. This made it difficult to give each person who listed as a grantor on the letter of attorney a unique identifier.  We resolved this imbalance by correlating grantors and grantees with places.  We equated grantors with their communities of origin and grantees with their place of residence, which means that the relationships of power of attorney are expressed in the maps and visualizations as connections between places.  We gave each letter of attorney a unique identifier as well.  We then created a relational database that connected grantors and grantees through the “event” of granting power of attorney (represented by the letter of attorney).

We focused on the following questions as we created the maps and visualizations:

1) What places (villages, towns, cities) represented “hotspots” for power of attorney (both granting and receiving)?  We decided to use CartoDB to express this.

2) What kinds of spatial/social relationships (expressed via lines between grantors and grantees) did power of attorney create (i.e. local, regional, colony-wide, transatlantic)?  We used Google Earth to express this.

3) How did the hotspots (question 1) and relationships (question 2) change over time?  We used an animated Gephi model to express this as well as static PDF maps for three different periods of time.

We produced the visualizations as follows:

Gephi. We used Gephi to build network graphs showing connections between places granting or receiving power of attorney. Network graphs require two types of data, nodes and edges. In our dataset, the nodes were places associated with the grantees and grantors while the edges corresponded to events when power was contractually transferred from one place to another. In other words, each edge represents the transfer of power of attorney from the location of the grantors to the location of each of the grantees list in a contract. Network graphs rely on different kinds of algorithmic layouts to make the connections more legible. In Gephi we used the GeoLayout algorithm to plot the nodes based on their geo-coordinates and a dynamic timeline plugin to animate the graph based on the date of the contracts. This allowed us to produce a visualization showing change over time and space.

Google Earth. We exported the network graphs created with Gephi to KML files, one for each time period. These files include the coordinate location of each place, as well as maintaining the network relationships. We were then able to open the files in Google Earth, overlaying the legal networks on detailed satellite imagery. The visualization emphasizes the spatial aspect of these relationships and makes clearer the physical separation between the mountainous towns of Villa Alta and the major cities they interacted with through the legal process.

CARTO. We used CARTO to create a series of interactive maps to orient researchers with the locations in our study, as well as to show ethnolinguistic regions, parish seats, and population change throughout the eighteenth century. We achieved this by creating a table of place names of all villages and towns involved in granting and receiving power of attorney and matching these with records in Geonames, giving us coordinates for each location. These places were then joined with other data such as counts of granting and receiving documents, population data, and political-administrative hierarchies among the towns. The ethnolinguistic region shapefiles were created by digitizing maps from historical literature on the region, geo-referencing them to a modern map of Mexico, and extracting traces of the regions as depicted in the eighteenth century.[2] The finished products of these maps allow researchers to explore population changes interactively and compare quantities of power of attorney letters granted and received in different towns. A map overlaid on satellite imagery also gives researchers a better sense of the mountainous terrain of Villa Alta and its geographical and topographical distance from the city of Oaxaca.