Now out in the International Journal of Lexicography, an article I’ve been working on for some time about the emergence of Twitter (as it was) and X (as it wishes to be) as the predominant source of new language evidence in the Oxford English Dictionary.
Here are some highlights and excerpts.
[Download PDF]When I say “new language evidence” I mean new language-evidence, as well as (and even more so) new-language evidence. Since 2018 or so Twitter/X has been the single biggest source of all evidentiary quotations in updates to the OED.
Obviously, as all Twitter quotes are dated 2006 or later, it is even more predominant (in fact, far and away dominant) when only modern quotations are considered.
Twitter/X also dominates various subcategories of language evidence added to the OED in the last ten years, including usage evidence of regional vocabulary (in virtually all regional categories), and of vocabulary marked for register (in all categories other than historical – though interestingly it is still the most prevalent of major sources when it comes to archaic words). 
Everyone loves the idea that someone could “invent” a word and have speakers adopt it. Shakespeare is said to have done this once or twice a day. On Twitter/X, this has happened, although for the most part OED’s first attestations of words from Twitter/X are taking advantage of a real-time transfer from the speech medium to the textual one. (This is most obvious in the case of “regional” first attestations; and least obvious perhaps in the case of the jargon particular to social media and computing).
Either way Twitter/X has been a major source of “new” words, both in proportional and in raw terms. Here’s an accounting, organized loosely into categories:
To me, among the most interesting and consequential of OED first attestations drawn from Twitter–which also may well be “twittergenic,” or actually first invented there–are their, them, themselves, and they to refer to . Generic gender-neutral they is ancient in English, but massive kerfuffles have occurred around this new and particular way of deploying it and its derivatives (specifically in reference to “a person whose sense of personal identity does not correspond to conventional sex and gender distinctions”) since the early twenty-teens:
“Long Covid” is another good candidate for OED-recognized Twitter word of the century (so far). This term emerged during the global lockdowns, as you, I and Oxford’s lexicographers were stuck working at home on our various connected devices.
Also isolated was Elisa Perego (@elisaperego78) who in the later Spring of 2020 was was tweeting daily within a dispersed, physically isolated community about long-lasting symptoms. On 15 May 2020 she replies to someone asking about “this ‘long tail’ form of Covid,” to say: “In Italy this is not much discussed for now. But the long covid is here, too, of course.” (https://web.archive.org/web/20200515195804/https://twitter.com/elisaperego78/status/1261345351820365824; also see Callard and Perego, “How patients made long Covid“). Soon the term spread to Facebook, and from there to the media, becoming well established in the discourse within a few months (compare “Long-haul Covid” or “long-term Covid”).
There are other neat examples of words for the times: shrinkflation, cryptocurrency, binge view, Brexiteer, remoaner, mansplainer, fat-shaming, deadnaming, to Zoom (v.) – these all appear to have been invented on Twitter, or at least first written down there.
The article talks about some advantages and some drawbacks to using Twitter/X as language evidence in this way. The main drawback is related to the main advantage, which is that Twitter/X is a very easy place to find all sorts of language–trillions of words have been “published” there since 2006–and so (like Shakespeare for early modern language evidence) there’s a danger of over-reliance. It is also, like some ancient language evidence, archivally precarious–I give an accounting of “tweet decay”, the disappearance of tweets over time, which seriously undermines the long term viability of the archive for historical linguistics.
No Comments