So you’ve got a new topic and a new project and you’ve no idea how to get started. And by you, I mean me, and by topic, let’s say specifically the ‘novel’ discipline of Digital Anthropology. For my first blog post of the academic year, I’m going to explore methodological challenges that appear in the initial exploratory phase of a project, and suggest some software tools that I’ve found beneficial. Specifically, I’m going to tackle the following two questions:
- Defining a Discipline: What exactly is “Digital Anthropology”?
- How to conduct a literature review? DH-Style. It has to be high-throughput and fast (which is to say I’m not manually clicking through 30,000 article hits in JSTOR)
My motivations for addressing these questions are pedagogical in nature just as much as they are derived from my own research needs. What makes me unique as a digital anthropologist, and what characteristics do I share with my colleagues the digital historians and digital scholars overall? How do I orient myself (and students) in this rapidly-emerging field and convey the significance of work that’s being done amidst traditional research values (is it real research if you’re using online data?). And finally, how can I quickly identify the most important literature to be engaging with so as to maximize my (and students) time and effort.
Defining a Discipline
What is Digital Anthropology? Until about a year ago, I’d never heard of this term as a discipline, despite my email inbox steadily filling up with conference calls from similar fields using the terms “Digital History” and “Digital Humanities”. I started out the way many researchers do, trying to find seminal works that will tackle this exact question in the introduction or foreword section. This will come as no surprise to digital scholars, as there is no consensus on an exact definition of the field. Is it traditional anthropological subjects and objectives, now just with digital methods? Is it studying the digital human in virtual worlds? Is it a method, or a new sub-field? The answer appears to be “yes” for all counts, and with all the controversy and conflict that comes with it.
Rather than pinning down a precise, sophisticated definition I am instead interested in broader questions concerning the relationship of digital anthropology, digital history, and digital humanities as a whole. What are the shared concepts and motivations that unite these disciplines? And what are the unique aspects that define the hazy disciplinary boundaries? To approach this question, I chose to do this in a fun/non-scientific manner by web-scraping summary webpages, most notably, Wikipedia.
Taking inspiration from Dr. Mica Jorgensen’s workshop Intro to Digital Scholarship (2018), I retrieved text from the Wikipedia page of each discipline (“digital anthropology”, “digital history”, “digital humanities”), and calculated word frequency using WordArt.com (Figure 1). I then selected the 20-most frequent words on each page and looked for both shared and unique words across the three disciplines (Figure 2).
So what are the shared concepts across all three disciplines? Unsurprisingly, the words “data”, “computer”, “university”, “research”, and “human” were the common themes that emerged (although “university” is probably being driven by the references cited section). But what makes digital anthropology unique? 5 words appeared: “community”, “social”, “virtual”, “ethnography”, and “ethics”. One may interpret this as suggesting that the discipline is foremost concerned with studying interactions between humans and communities in a virtual world via ethnography (with an eye out for ethics, enough for it to have a dedicated section). But in the end, what does Wikipedia know?
A very simplistic benefit of this fun exercise is an improvement to my key word search phrases. From the word frequency charts and context, I discovered that “digital anthropology” also goes by the following 3 synonyms: virtual anthropology, technoanthropology, and my new favorite: cyberanthropology. And thus I can now construct more comprehensive keyword searches that account for discipline and regionally specific language variation. And speaking of literature keyword searches…
Literature Review: DH-Style
Instead of relying on a Wikipedia summary, I turn to the more robust form of knowledge acquisition: the comprehensive literature review. But sifting through 30,000 google scholar hits neither appeals to me nor is it actually a feasible option. The fact is that online literature repositories have fundamentally changed the ways in which we access and engage with the research literature. While a deluge of data may be appealing, it has the undesirable side-effect of sometimes making matters worse, by “calling our attention to the inconvenient fact that an impossibly large number of sources could be read and evaluated before we write” (Edelstein, 2014, pg. 2). If it’s infeasible to comprehensively consult all relevant literature, how are we to have confidence that we are engaging with appropriate arguments and theories? One solution: if powerful digital tools create this problem by generating big data, let’s match it with a digital tool of equal power in analyzing voluminous information. Let’s go big data, let’s go high-throughput, let’s go with a DH-style of lit reviews.
At this point I am exploring the use of two bibliometrics/data-mining approaches for conducting a lit review: dimensions.ai and JSTOR’s ‘Data for Research’ service. First is the web-tools dimensions.ai (Figure 3) which provides exploratory stats and literature links for a search query. I was interested in what range of dates I should begin consulting for publications and noticed two relevant trends (Figure 4). The first is that something ‘interesting’ happens in 2014 where publications abruptly double in number and then recede. The second is that citations are constantly increasing, which affirms to me that there is an active and growing interest in using theories and methods within digital anthropology. My planned use for this tool is to both explore and refine relevant literature that can provide novel theories and interesting (and hopefully publically accessible) datasets for novel analysis.
The second web-tool that I’m learning right now is JSTOR’s Data for Research (DfR) service which provides all retrievable metadata (and n-grams) from a particular search query. My “digital anthropology” search query returned over 25,000 hits and the metadata file was 1.3 GB in compressed zip form. Data acquisition is extremely easy, but processing it is where things get challenging. While there are many excellent R packages provided to accomplish this task, I would prefer a tool that does not have such a steep learning curve so as to make this form of analysis possible for students to learn within a workshop or academic term. To find a more appropriate tool, I’m going to be sifting through Andy Kirk’s amazing but immense software visualization repository (Figure 5). Until next time!
Tools and Links
Dimensions. Literature Review Exploration. https://www.dimensions.ai/
Palladio. Visualizing historical data with ease. https://hdlab.stanford.edu/palladio/
Software Visualization Repository. http://www.visualisingdata.com/resources/
WordArt. WordFrequency. https://wordart.com/
Edelstein, D. (2014). Enlightenment scholarship by the numbers: dfr.jstor.org, dirty quantification, and the future of the lit review. Republic of Letters, 4(1), 1-26
Jorgensen, M. (2018). Intro to Digital Scholarship. Presentation at the Do More With Digital Scholarship (DMDS) Workshop Series. Hamilton, ON.