It’s dissertation season in the Faculty I work in, which means it’s a time of referencing questions a-go-go. Like most things, referencing is a mix of common sense, cobbling something together that looks roughly OK, and being consistent about it. In the past three days I’ve been asked about referencing sourceless, orphan works found in random bits of the internet, live dance performances from the early 20th century, and – in another worlds collide moment – how to reference an algorithm.
A student was basing a portion of their argument on the results of Google’s autocomplete function – this kind of thing:
My colleague and I were stumped. Who owns this algorithm? Well, Google. But it’s also collectively formed, discursively constituted, mutually produced. How do you reference something that is a temporary, unstable representation?
Pickering (1993, 2002) argues that ‘things’ move between being socially constructed via discourse and existing as real, material entities – a performativity which is “temporally emergent in practice” (p. 565), a kind of mangled practice of human and material agency which emerges in real time. This kind of autocomplete text (if ‘text’ is the right word) reflects this completely.
The act of referencing is one of stabilising, as well as avoiding plagiarism or practising academic integrity. When referencing online sources which don’t have a DOI or a stable URL, you are artificially fixing the location of something and representing it via text. You put ‘accessed’ dates to secure oneself against future accusations of plagiarism but also in view of the instability of the digital text. It’s not an ideal process, but it works.
And yet referencing – or indicating ownership of an autocomplete algorithm – seems to take this a step further. It leans towards reification, and it imbues the algorithm with a human and material intentionality which isn’t justified. It ‘essentialises’ what is fleeting and performative. So how, then, do you capture something which is, as Pickering writes it, ‘temporally emergent in practice?’
I suppose I should say what we told the student too, though it may not be right. We suggested that it didn’t need to be referenced, because it constituted their ‘own’ research; you wouldn’t reference the ‘act’ of reading, or the technology used to find, access or cite resources. You’d cite someone else’s published ‘version’ of the algorithm, but not your own. This uncovers another area where digital technology shapes and is shaped by ‘traditional’ practices and performances.