Brits are people, citizens are from abroad and foreigners will kill you

Christophe and Cory address the idea of a citizens' assembly and analyse the ways that different people respond to the word "citizen".

The idea of a citizens’ assembly has been proposed as a way of bringing a divided country together.

Our contribution to this proposal is to analyse the conditions under which this civil society exercise could take place. We are conducting a large-scale analysis of language use in the UK (and Ireland) today, focusing on the term “citizen” and related concepts around social and political status, and are able to present our initial findings here today.

While the word “citizen” once covered a wide range of activities and interests, including a person’s contribution to society at large and the way in which their views were represented, public discourse in the UK over the past several months has narrowed considerably, and largely restricts the term to one specific group of people only. The results are disturbing, the picture being one of crisis, strife and mistrust. Deep divisions between Remain and Leave leaning media can be observed with regard to every single term and phrase we have examined.

In order to find out how people in the UK refer to each other we are employing an innovative research method called word2vec. Rather than looking at which words are used in any given medium and what contexts they appear in our method is to ask which words are most similar in the way they are being used. Simply put, we are not trying to find out which words are often used together, but which ones could stand in for one another.

We do this on the basis of a dataset of about 16,000 articles from British and Irish newspaper and political campaign websites, comprising roughly equal numbers of Remain and Leave leaning outlets. Word2vec is designed to test pairs of phrases against the dataset in order to predict which word is most likely to appear between them. We assume that readers, when seeing the two phrases, will anticipate that word. In addition, we can compile a list of other words that are likely to appear in the same place, albeit ever so slightly less so. These can count as synonyms, i.e. words with nearly the same meaning, as they are used in roughly the same context. So while traditionally linguists have tried to derive the meaning of a word from the context in which it is used, we try and predict which words will be used by looking at existing phrases. Using this method, we learn more about a word’s meaning and, crucially, about the attitude of readers.

The results obtained paint a picture of deep divisions: “Citizens” now refers primarily to non-UK nationals, notably Irish nationals, and is used in similar ways as negatives like “undocumented” and “false identity.” This stands in sharp contrast to established notions of citizenship, which tend to focus on those who are part of a given polity.
When we look at the term “British,” we find a prevalence of terms indicating that they are concerned with making their “voices heard” as they ask “who cares about” them. “Foreigners”, meanwhile, have become near synonymous with violent Muslims who are a drain on taxpayers’ money and cause public disturbances – the words most similar to “foreign” include “local authorities,” “hospitals,” “councils,” and “prisons” as well as “rough sleepers,” “modern slavery,” and “knife crime”. These are the findings from the complete dataset, comprising both Leave and Remain supporting media and campaign websites.

Looking at the two political camps separately, vast differences in outlook emerge. On the Remain side, “citizen” is akin to terms surrounding legal process and political activity (“ruled,” “tribunal,” “dispute”; “activities,” “protesters”) whereas among Leavers, the word is more emotionally charged (“crimes,” “victims,” “bombings,” “explosion,” “suicide bombers,” “death penalty”).

As to the term “British,” two vastly different visions of the people of this country are in evidence. For Remain – apparently in the wake of the Brexit referendum and the sequent (lack of) political moves – the term “British” bears similarity to a litany of negative terms (“unacceptable,” “disaster,” “no clarity,” “catastrophe,” and the wonderfully colourful “shambolic”) as well as some that relate to compromise or blockade (“veto,” “negotiators,” “both sides,” “long extension”). For Leave, the term “British“ is used with a note of tribalist triumph: “trust,” “strong,” “most important,” “undoubtedly,” “pragmatism,” and “spirit” are all used in similar ways.

Crucially, the two camps are so different that any specificity vanishes from the results when we combine them. In the overall dataset, “British” means “us,” and it marks a place and a people – but that place and people are either a shambolic ruin or a strong union, depending on who you ask.

It seems to us that what we have presented here provides a strong indication that, in setting up listening and deliberation exercises, we need to take account of deeply entrenched and widely diverging ways of talking about one another.

Christophe and Cory are continuing their analysis, and are happy to discuss both the findings presented here and possible directions of future research.