I learned last month from Wired that something along the lines of what I’ve been calling “algorithmic culture” already has a name — culturomics.
According to Jonathan Keats, author of the magazine’s monthly “Jargon Watch” section, culturomics refers to “the study of memes and cultural trends using high-throughput quantitative analysis of books.” The term was first noted in another Wired article, published last December, which reported on a study using Google books to track historical, or “evolutionary,” trends in language. Interestingly, the study wasn’t published in a humanities journal. It appeared in Science.
The researchers behind culturomics have also launched a website allowing you to search the Google book database for keywords and phrases, to “see how [their] usage frequency has been changing throughout the past few centuries.” They follow up by calling the service “addictive.”
Culturomics weds “culture” to the suffix “-nomos,” the anchor for words like economics, genomics, astronomy, physiognomy, and so forth. “-Nomos” can refer either to “the distribution of things” or, more specifically, to a “worldview.” In this sense culturomics refers to the distribution of language resources (words) in the extant published literature of some period and the types of frameworks for understanding those resources embody.
I must confess to being intrigued by culturomics, however much I find the term to be clunky. My initial work on algorithmic culture tracks language changes in and around three keywords — information, crowd, and algorithm, in the spirit of Raymond Williams’ Culture and Society — and has given me a new appreciation for both the sociality of language and its capacity for transformation. Methodologically culturomics seems, well, right, and I’ll be intrigued to see what a search for my keywords on the website might yield.
Having said that, I still want to hold onto the idea of algorithmic culture. I prefer the term because it places the algorithm center-stage rather than allowing it to recede into the background, as does culturomics. Algorithmic culture encourages us to see computational process not as a window onto the world but as an instrument of order and authoritative decision making. The point of algorithmic culture, both terminologically and methodologically, is to help us understand the politics of algorithms and thus to approach them and the work they do more circumspectly, even critically.
I should mention, by the way, that this is increasingly how I’ve come to understand the so-called “digital humanities.” The digital humanities aren’t just about doing traditional humanities work on digital objects, nor are they only about making the shift in humanities publishing from analog to digital platforms. Instead the digital humanities, if there is such a thing, should focus on the ways in which the work of culture is increasingly delegated to computational process and, more importantly, the political consequences that follow from our doing so.
And this is the major difference, I suppose, between an interest in the distribution of language resources — culturomics — and a concern for the politics of the systems we use to understand those distributions — algorithmic culture.