The Musings of a Contemplative Woman

“Historically, counting and classification have been used to dominate, discipline, and exclude.”¹

This week (or should I say last week?), I think I was left with more questions than answers. Maybe because I didn’t have enough time to stew on everything I did thanks to pre-reading week assignments, but I’d like to think that it’s because my brain is working on trying to grasp a concept beyond my knowledge, and is waiting to learn more.

Central to this week was the concept of “what gets counted, counts,” and in reading the article of the same name by D’Ignazio and Klein, I can see it applied to almost every structured system I’ve encountered. There’s never room for ambiguity in the archive or databases; even in spaces where it’s desperately needed, when deciding between allowing ambiguity or placing restrictions at the expense of accuracy, the latter is almost always selected because of the inherent belief that data must be neatly categorical and organized as such. This concept converges with another notion posited in the article by Labrador where she discussed the cataloguing of artefacts and how many often only support a single form of identification, and I while reading that I distinctly recalled the first time I acknowledged this restriction when it came to the methodology of how databases are created and standardized. I was working on a project cataloguing heritage items within a heritage architecture context. Much like how we documented our cemetery, there was a numbering system created specifically for this project, and a database to hold each item after it was categorized. Yet even with a system specifically designed for the project, I still recall facing the challenge of “where does this go?” Many objects were simple to categorize, like wooden a floor plank, but then something like a stained glass window with attached iron patterning and framing would come up. Windows was a category, but when inputting materials, only one material could be selected. These windows were much more than just “glass”, but the choice I had was to either select that option or leave the form blank, which really, was not an option because if I wanted to keep my job I couldn’t just keep skipping fields in the database I was adding to.

This lack of room for uncertainty once again reappeared in a more powerful way while documenting my chosen cemetery for our class graveyard project. Even before the process of “proper” documentation via Kobotoolbox began, when I was just out in the field drawing a map of the cemetery at Pinhey’s Point, taking pictures, measuring, I thought to myself many times, “Wow, I know virtually nothing about the lives of these people.” When it came time to enter what I recorded into Kobotoolbox, I couldn’t help but think, so much of what I’m entering is assumptions– assumptions about gender, about fonts, about the year a person was buried. And further, after reading the D’Ignazio and Klein article, I started reflecting on that which I didn’t count. I stated in an earlier post that I officially recorded all of the graves in the cemetery with a death date of 1905 or earlier, but according to Meg, the kind cemetery keeper and 6th generation of the family that established the cemetery, the founder of Pinhey’s Point had a favoured slave named Abdul who upon death requested to be buried in the family’s cemetery. Because Abdul was not Christain, he was buried as close to the cemetery as possible yet could not be buried within it. Years later the cemetery was expanded to accommodate the growing church community, and with this expansion, it is now believed that Abdul was granted his wish of being buried in the walls of the otherwise closed off cemetery. Abdul’s grave was not marked– or at least, not marked using a material that lasted– and thus in my survey of the cemetery at Pinhey’s Point, despite having died prior to 1905, he was not counted. Hearing this story made me wonder how many others may have been buried in that cemetery but not thought of as significant enough to be memorialized and due to this left entirely unaccounted for as those who remembered them died as well.

With all this brewing in my mind, I was eager to start this week’s work, specifically the activity in which we were given the opportunity to alter the Kobotoolbox form used for recording the cemetery data. I thought, without specificities, that I could definitely improve it; yet, when it came to looking over the form the first thing I realized was that I had no idea how I could make the form better. Perhaps it’s how new I am to the field of archaeology, but I couldn’t think of how I could change or add any questions to the form to improve the relatively impersonal data that came from it. When it came to the form itself, I primarily had problems with the entry-type restrictions imposed on each question. I couldn’t enter newlines when recording the inscriptions on each grave, making the inscriptions less readable when typed and not allowing me to preserve the form of the message in a recognizable way. More importantly, the “age” boxes were restricted to only a numerical value. While age is typically recorded numerically, this became a problem when it came to the deaths of children. There were a number of times where I had to enter the age in the “additional notes” section because due to the box allowing only whole numbers, I couldn’t specify in any way that the number “3” I was entering in the box represented months rather than years. The age of these children could not be recorded in the category designated as being for the age of those buried, and further, their age was left as a note that may be ignored or that creates a flaw in the dataset, potentially skewing the results of manipulations with the blank space it leaves.

What I mistook as struggling with the Kobotoolbox form was actually the issues of understanding I had with the DEBS coding scheme itself. These problems were primarily rooted semantics, both when it came to translating aspects of UK English to Canadian, and lack of knowledge surrounding archaeological terminology. There were a couple of categories I’m sure there is an error or that I consciously left blank because not only did I not quite understand what I was being asked to record, but also I questioned their relation to Canadian memorialization practice. Luckily, a majority of the headstones were evidently made from the material marble, ever-popular in 19th century headstone masonry, but there were also several graves made from stone that I couldn’t identify, leaving me staring at my photos caught in a loop of contemplating slate, limestone, or maybe rough granite. And I stuck to these stones because they were the only stone varieties listed on the code sheet that I was sure were types of stone that could be found in Canada. In the same vein, I wasn’t sure about tooling techniques, and when I searched for a guide all I could find was a guide on modern tools for tooling! On top of not quite understanding what tooling is, I also didn’t know if the same technique that were used in the UK translated over to Canada, where a new variant of the technique may have developed. This is the type of category I consciously chose to leave blank.

During this Kobotoolbox activity combined with the peek into the dataset created by our class inputting the graves they recorded I saw in the intro to stats lesson in R, the second thing I realized was the results of restriction when documenting a subject so full of uncertainty. There were numerous holes in the data, often at least one in each category, revealing to me that I wasn’t the only person facing an issue of understanding and ambiguity of the unknown. Reflection on this new revelation of collective doubt is where most of my questions began and have yet to find an end. Going through the columns of what and wasn’t documented, I thought, while we were not those who created the form we used to record our findings, through the data we captured we still fundamentally, collectively constructed this dataset through additions and omissions. Yet the fact that the dataset has omissions indicates that there were elements of the form that either were not understood or that caused those entering the data to feel stuck should they not have a clear answer to enter. But then how does one create a form that can accommodate the ambiguity that inherently comes from the study of human civilization– human life? And further, how can something like a dataset which is so consistently considered a representation of order and categorical information be reworked to be something usable yet permissive of things that are beyond categorization? Lastly, it’s harder to analyze datasets which allow for ambiguity through digital methods, so how do we, as digital humanists adapt our methods of analysis to this? While we must always think critically of the results our digital methods produce and, further, the technology itself, how do we adapt this technology to become less structured, able to handle variability, become more human?

¹“What gets counted, counts”. Chapter from Data Feminism by Catherine D’Ignazio and Lauren Klein

Picture featured is screenshot from Codex Atlanticus