Popular imaginative geographies and Brexit: Evidence from Mass Observation

The EU Referendum of 2016 was one of the most significant events in recent British political history. It is widely recognised that citizens engaged with the referendum through understandings of Britain, the EU, the world, and their place in it. This study complements existing research where such understandings have been inferred from citizens ’ demographic characteristics, the characteristics of their localities/regions, or elite discourses. It builds on existing research where a more direct engagement with citizens ’ understandings has been achieved through interviews or focus groups, allowing the content of understandings to be thickly described. To these latter studies, this paper makes three main contributions. First, it focuses on popular imaginative geographies, which are conceptualised drawing on literatures in Geography and Political Science as fast ‐ thinking heuristics. Second, it brings new evidence to the conversation in the form of volunteer writing for Mass Observation. Third, the focus is on the content of popular imaginative geographies, but also how and why such geographies were used by voters in the referendum. The main findings include that many Leave supporters imagined Britain as an island – either a once great military and imperial power, an island separate from Europe, needing freedom from Europe to engage in the wider world; or a small island, a full container, close to the rest of Europe and vulnerable to mobilities across Europe ’ s borders. By contrast, many Remain supporters imagined Britain as post ‐ imperial, small, vulnerable, and under threat of isolation from Europe and exposure to a chaotic, uncertain, dangerous world. Both groups engaged with the referendum through such popular imaginative geographies because the referendum presented voters with a difficult task, the campaigns provided few trustworthy


| INTRODUCTION
Since June 2016, when a narrow majority of eligible British citizens voted to leave the European Union, geographers have engaged with Brexit -British exit from the EUin a variety of ways. The potential consequences of Brexit have been considered for different regions of the UK (Billing et al., 2019), different sectors of the UK economy (Brown et al., 2019;Dobruszkes, 2019;Maye et al., 2018), and different financial centres across Europe (Dorry, 2017;Lavery et al., 2018); the way Brexit has been variously felt and enacted as an event since the referendum has been considered (Anderson et al., 2020), as have the experiences of certain groups of citizens, including non-British citizens living in the UK (Botterill, 2018;Botterill et al., 2018), non-white British citizens living in the UK (Redclift & Rajina, 2019), and British citizens living in other EU Member States (Higgins, 2019;Miller, 2019). In the present paper, we consider British citizens living in the UK, their understandings of Brexit and related categories like Britain, the EU, and the worldwhat we'll term "popular imaginative geographies"and how these understandings shaped the referendum vote.
As a complement to such studies, in this paper we seek a more direct engagement with citizens' understandings. We try listening to citizens when they write about Brexit, Britain, the EU, and the world in their own terms. To date, there have been only a handful of other studies that have attempted such a direct engagement (Anderson et al., 2020;Andreouli & Nicholson, 2018;Bromley-Davenport et al., 2019;Willett et al., 2019). The present paper builds on these existing studies, focused on qualitative data and thickly describing the content of understandings, in three main ways. First, we focus in particular on the popular imaginative geographies of Leave and Remain voters, which we conceptualise drawing on behavioural Political Science as fast-thinking heuristics (Section 2). Second, we bring a new dataset to this particular area of research: volunteer writing for Mass Observation (Section 3). With one exception (Moss et al., 2020), this dataset has not yet been used to study Brexit, despite having been used successfully in recent research on popular understandings of politics more broadly (Clarke et al., 2017(Clarke et al., , 2018 and also classic research on British attitudes to Europe, Europeans, self, and other (Kushner, 2004). Third, we not only describe the content of these popular imaginative geographies (Section 4), but also demonstrate how and why they came to be used by voters in the referendum (Section 5).

| POPULAR IMAGINATIVE GEOGRAPHIES AS FAST-THINKING HEURISTICS
We take the term "popular imaginative geographies" from Gregory (1994). Influenced by Said (1978), among others, Gregory identifies not one geographical imaginationa sensibility recognising the importance of space and place in the conduct and constitution of social lifebut multiple imaginative geographies: accounts of the inscription of social life in space that are always local, particular, situated, positioned. These partial accounts need interrogating and enlarging. They get produced when people seek to make sense of their own and other cultures and landscapes through writing, painting, and photography. They exist as repertoires of categories, codes, and conventions that shape practices. They reflect the gaze of elites, but also circulate widely in society as popular imaginative geographies, functioning as common-sense understandings of places and spaceshow they differ and relatewith material effects on everyday life.
We use the term "imaginative geographies," but draw from literatures using other terms too. These include writings on "spatial imaginaries" (see Watkins, 2015). They include writings on "popular geopolitics"at least where that term is used to describe geopolitical understandings found not only in cultural products like newspapers and films, but also among the audiences for such cultural products (see Dittmer & Dodds, 2008). They also include writings on "mental maps"at least where that term is used metaphorically to describe geographical understandings in the same broad sense meant by imaginative geographies, and not only in the narrower sense of image-plans sketched by research subjects (see Tuan, 1975).
All this provides a number of potential research questions for a study of popular imaginative geographies and Brexit. What representations of Britain, the EU, the world, and Brexit were held by British citizens around the time of the 2016 referendum? How did they vary between individuals and groups of citizens? How did they vary from material geographies? From where did they originate? And how did they inform citizen decision-making? Our focus in the study from which this paper draws was on the first and last of these questions. Our dataset and study design did not allow for patterns to be identified below the level of Leave and Remain voters (e.g., differences between the imaginative geographies of working-class and middle-class Leave voters). Nor did they allow us to establish the origins of these popular imaginative geographies, likely to be found in the material conditions experienced by certain groups of citizens and also the framing effects of elite discourses. Instead, the affordances of the Mass Observation (MO) data, discussed further in Section 3, encouraged a focus on the representations held by citizens, how they varied between Leave and Remain voters, and how they informed the decision-making of these voters around the time of the referendum. These are crucial parts of a full explanatory account of the referendum vote, as we note in the concluding section of this paper.
To conclude the current section, we have one final set of points to make about popular imaginative geographies. Reading the literature, we were struck by the similarities in much writing on imaginative geographies, spatial imaginaries, popular geopolitics, and mental mapsdetailed previously in this sectiondespite the different paradigms from which some of these concepts originated (Johnston, 1997) and the differences emphasised by others (Watkins, 2015). We were also struck by the similarities in much of this writing by geographers with literatures in the wider social and human sciences -Psychology, Economics, and especially Political Scienceon cognition, reasoning, and (political) action.
The view from behavioural Political Science is that citizens do not pay much attention to politics (Marcus et al., 2000;Zaller, 1992). They are minimally informed about politics (Delli Carpini & Keeter, 1996). Because of this, citizens make judgements about politics using "System 1" thinking (Stanovich & West, 2000) or "fast" thinking (Kahneman, 2011): intuitive, emotional, automatic thinking, as opposed to analytical, controlled, sequential thinking ("System 2" or "slow thinking"). Fast thinking requires a model of the world, of what is normal, providing cognitive cues or heuristics for understanding, justification, and action (Kahneman, 2011). Such models are made up of categories, cases, types, and exemplars (Lakoff, 2002). They are shaped by everyday experience, social interaction, expert knowledge, and institutional environments (Kempton, 1986;Lupia & McCubbins, 2000). As such, they are shared by social groups. They are also less technical, less coherent, and more contradictory (more happily contradictory) than formal theories or political ideologies (Keesing, 1987;Lakoff, 2002). Nevertheless, they are used by citizens to guide behaviour because they work well enough as shortcuts in most situations (Keesing, 1987).
Not all of these positions can be found in writing on imaginative geographies. Nevertheless, in behavioural Political Science, to understand the political judgement and action of citizens is to identify their cultural models (and the categories, cases, types, and exemplars from which they are made). In the present study, to understand the views and votes of British citizens regarding Brexitwhich is partly a question about Britain and the EU, self and other, and the relations between these categoriesis to identify their imaginative geographies (and the images and narratives from which they are made). We have therefore developed the concept of popular imaginative geographies by drawing on literatures in Geography but also the wider social and human sciences. In this paper, we view popular imaginative geographies as System-1 or fastthinking heuristics that inform the political decision-making of citizens alongside other cues and shortcuts, e.g., the elite cues commonly thought to be important in referendum voting (Hobolt, 2006(Hobolt, , 2007. In the rest of this paper, we identify the popular imaginative geographies shaping the views and votes of British citizens regarding Brexit in volunteer writing for Mass Observation.

| MASS OBSERVATION AND BREXIT
Mass Observation (MO) was founded in 1937 to record the everyday lives of ordinary people in Britain and especially to enable the masses to speak for themselves and make themselves heard above the voices of the press and politicians claiming to speak in their name (Hinton, 2013a). It was established as a complement to polling by the British Institute of Public Opinion, also founded in 1937, and MO collected a variety of material by a variety of meansnot only tightly worded survey questions but also open-ended (sets of) questions known as "directives," sent every three or four months to a panel of volunteer writers. The first incarnation of this panel existed from 1939 to 1955. MO as a whole ceased activity in the mid-1960s. In 1975, the Mass Observation Archive was opened at the University of Sussex. In 1981, MO's panel of volunteer writers was re-established by the Archive as the Mass Observation Project. Ever since then, directives have been sent every three or four months to approximately 500 volunteer writers. In 2016 and again in 2017, panellists were asked to write about the EU Referendum and Brexit. The full wording of these two directives can be found in Table S1. Responses to these directives provide the dataset for the present study.
MO materials have been used by many historians and social scientists (Hinton, 2010;Kushner, 2004;Kynaston, 2007;Langhamer, 2013;Savage, 2010). Still, there are some methodological issues that any study using volunteer writing for MO should address, including representation on the panel and how the material should be approached and analysed. The original MO panel has been criticised for its social constitutionfor over-representing the radicalised lower middle class (Jeffrey, 1978) or the technically minded middle class (Savage, 2008), along with people from London and the South East. The 21st-century panel of the Mass Observation Project has received less criticism of this kind, largely because it has been more carefully constituted following the norms of contemporary social research. Nevertheless, we sampled within the panel, following the example of Salter (2010) and using MO's database of panellist characteristics, supplemented by details found in the responses themselves. The purpose of our sampling strategy, given the qualitative character of the data, was to identify the full range of popular imaginative geographies and so to achieve descriptive saturation (Baker & Edwards, 2012). This was achieved by sampling 60 respondents for each of the two directives and filling quotas for age group, gender, country/region of the UK, occupational category, and affiliation to Leave/Remain.
The full details of the final sample can be found in Table S2. The gender split was roughly equal, as was the split between Leave and Remain voters. All age categories were covered, though categories towards the middle of the range (40-49, 50-59, 60-69, 70-79) were a little better covered than categories towards the two ends (18-29, 30-39, 80-89, 90-99). All countries/regions of the UK were covered, with the exception of Northern Ireland, which is poorly covered by the full MO panel. The most challenging characteristic to sample for was occupational category, again because some categories remain poorly covered in the full MO panel. Compared with figures for the UK population as a whole (ONS, 2018), "professional" and "administrative and secretarial" occupations were over-represented, while "skilled trades," "process, plant, and machine operatives," and "elementary" occupations were under-represented.
We sought to correct for some of these remaining limitations of the sample by how we analysed the data and wrote up the findings. We analysed the data by coding the writing of MO panellists for popular imaginative geographies, the categories and storylines from which they are made, and how they were used. Tally tables were then used to check for patterns and especially relationships between themes (certain popular imaginative geographies, categories, storylines, modes of use) and the demographic and other characteristics of panellists (gender, age group, occupational category, and so on). The main significant difference found was between Leave voters (who tended to be olderreflecting what is already known about Leave voters) and Remain voters (who tended to be younger). This gave us the present paper's focus and structure. It may be that other differences existe.g., between middle-class and working-class Leave votersbut such finer-grained differences did not show up clearly in the MO data.
In writing up the findings, we sought to include as many different voices as possible and not to privilege or silence particular groups. Again, the outcome of this effort can be seen in Table S2. Roughly equal numbers of Leave and Remain voters are quoted in the paper, though some Leave voters are quoted multiple times. Slightly more women than men are quoted, with some women also quoted multiple times. Slightly fewer panellists in the 80-89 and 90-99 age groups are quoted compared with panellists in the other age groups. Reflecting the limitations of the sample, more "professionals" and panellists from "associate and secretarial" occupations are quoted compared with panellists in other occupational categories, though some panellists in other occupational categories are quoted multiple times. In terms of country/region of the UK, the numbers are small and should not be over-interpreted, but they are smallest for regions in southern England (London, the South East, and the South West) and also Wales and Northern Ireland.
Ultimately, we tried hard to give voice to a broad range of panellists and largely succeeded in doing so, though we acknowledge that some voices may still be amplified a little in the paper compared with others. Moreover, we argue that issues of representation should not be over-emphasised when dealing with volunteer writing for MO. There are different possible approaches to analysing MO data. One could see panellists as representatives of particular social groups (e.g., women or young people), but the writing of panellists can also be read for the access it provides to discourses, or elements of discourse, circulating in society at a particular historical momentespecially where multiple panellists, from a variety of social positions, may use similar categories or storylines in their writing. This latter approach draws from cognitive science (Holland & Quinn, 1987;Lakoff & Johnson, 2003), which sees behaviour as shaped by understanding, and understanding as shaped by cultural knowledge in the form of shared models, schemas, frameworks, scripts, stories, metaphors, and prototypes. It draws from social theoretical writing on discourse (Foucault, 1991), which sees social reality as shaped by practices, and practices as shaped by forms of consciousness made up of concepts, ideas, representations, images, frames, stories, and narratives. It also draws from interpretive social science (Bevir & Rhodes, 2005), which sees action as shaped by beliefs, and beliefs as holistic, such that action takes place within discourses or traditions. This "horizontal" approach (Clarke et al., 2018) has often been taken to MO sources, especially in the discipline of History. By reading the referendum diaries for cultural resources shared by large groups of panellistsin our case, Leave and Remain voterswe followed the examples of: Salter (2010), who read MO sources for "shared" or "public" understandings; Nettleton and Uprichard (2011), who read them for shared "cultural repertoires"; Gazeley and Langhamer (2013), who read them for shared proverbs, truisms, and "everyday episteme"; and Hinton (2013b), who views MO panellists as particularly reflective people who, precisely because of this quality, provide researchers with access to the cultural worldthe world of discoursethey and others inhabit.
In brief, then, the study aimed to establish the range of popular imaginative geographies available to voters around the time of the referendum, their content, their adoption by Leave and Remain supporters, and how they came to be used by voters in the referendum. Despite their limitations, we argue the sample and selected quotes are sufficient for these purposes, which in turn are appropriate to the relatively small-n, qualitative dataset we analysed. We structure our discussion of findings around two "where" questions.

| WHERE IS BRITAIN?
Remain supporters tended to imagine Britain as one among many small European countries that need the EU because of their small size and consequent vulnerability. A community health worker in her 40s from the East Midlands (T4715) wrote: The UK is a tiny place. The relative safety we've enjoyed over the past 30 years since the Cold War ended feels under threat at the moment. Some of the tensions occurring now feel similar to the tensions that started the First and Second World Wars. Surely it would be better to be a part of a bigger organisation in order to deal with this?
Similarly, a special education needs support assistant in her 40s from Yorkshire and Humber (B5178) wrote: "We are relatively small nations in Europe in comparison with the US, China, and Russia, and it makes sense to me that we unite to trade and allow free movement of goods and people." If Britain was "tiny," "relatively small," and "under threat," then for many Remain supporters this was partly because Britain was no longer the great imperial power it once was. For a caseworker in her 50s from Wales (J2891), "We are a former empire with nothing left. We cannot feed ourselves. We have no industry left. The government have sold off all assets so that we are mostly foreign-owned." A saleswoman in her 20s from London put it succinctly and forcefully (M5578): "We are not a bloody 19th-century empire anymore!" A self-employed professional in her 40s from the North West (G226) justified her Remain vote as follows: "I can see lots of advantages of being part of something larger. I feel that Britain is no longer 'great.'" A pharmacist in her 50s from the West Midlands (V3773) justified her Remain vote in similar terms: I voted Remain because I thought that nations needed to cooperate, learn, and grow together. I thought it was arrogant to expect Britain to be great again on her own. When Britain was 'great,' we controlled large areas of the world and used other people's resources, and that won't and shouldn't happen again.
For this panellist, Britain was no longer "great" and perhaps never had been great by some important measures. Remain supporters more generally shared images of Britain as not only small, but also getting smaller; no longer great, but in great need of European unity.
An important category for Remain supporters, repeated by many panellists in response to both directives, was "isolation" and the potential for Britain to be "isolated" in the world if it left the EU. Before the referendum, a student in his 20s from the North West (S5780) expressed his "greatest fear": "that a vote to leave triggers years of uncertainty and leaves us isolated with a weakened negotiating position on the world stage." After the referendum, a teacher in his 40s from the South East (P5366) worried that, because of Brexit, "We will become very isolated as a country." Another teacher, this one in his 30s and from the North West (P5715), justified his Remain vote as follows: "Living in isolation from countries with whom we can share so much makes absolutely no sense." For many panellists, the figure of isolated Britain was connected to the figure of small, post-imperial Britain. "We are leaving ourselves very exposed without Europe," wrote a sales ledger controller from Northern Ireland (C5692). "The UK isn't a big power anymore … We should be strengthening the bonds with other countries and not leaving ourselves out in the cold." This figure of isolated Britain was also connected to another figure: a wider world increasingly characterised by uncertainty, danger, and chaos. Listen again to the teacher in his 30s from the North West (P5715): "In this increasingly uncertain world, with the threat of Islamic extremism coming from groups like ISIS, we need community and partnership more than ever." Or take this from a retired finance officer in her 60s from Scotland (F5890): "I feel we would be much better as part of a large European community than standing alone against a growing tide of terrorism and world unrest." A film writer in his 70s, also retired and from Scotland (H1541), demanded that "we remain and work together," because "The world is such a chaotic and potentially dangerous place … common sense demands nothing less." Remain supporters, it would seem, shared an imaginative geography populated by the following central figures: small, postimperial, vulnerable Britain; under threat of isolation from Europe and exposure to a chaotic, dangerous, uncertain world. How does this compare with the imaginative geographies of Leave supporters? When writing about Britain, Leave supporters tended to write about an island separate from "continental" or "mainland" Europe. "I have never felt personally European, always British," wrote a teacher in her 60s from Scotland (W729). "Britain has always been apart from Europe, geographically and culturally, and all this integrating with Europe is just nonsense." A market researcher in his 60s from the North West (P4287) wrote: "We are a free country with our own needs and direction. We have a different outlook on the world as an island. We have historically taken a different path than continental Europe." This historical narrative of Britain's development as an island separate from the rest of Europe was also told by a retired nurse in her 80s from the East Midlands (M2061): "We have been an Island Nation for thousands of years. Why are we now linked to Mainland Europe by an 'invisible rope'?" For many Leave supporters, this narrative was closely associated with a story of Britain as a formerly great, and potentially still great, military and imperial power. The retired nurse from the East Midlands (M2061) again: "This is a Great Country. We have survived two world wars … Our brave young men did not fight for us to be ruled by Europe." Or listen to this administrative assistant in her 20s, also from the East Midlands (S6115): "We are historically a strong and robust set of people. We've fought wars and communities have grown and thrived through altering conditions." The idea here is that Britain can stand alone now because of its "great" military history. The idea is also that Britain can stand separately from Europe now because of its "great" imperial history. We see this in frequent mentions of "the Commonwealth" and its members by Leave supporters. A woman in her 80s from the East of England (H260) wrote: "I was raised on a diet of patriotism, especially during the war years. Then, our country traded with … Canada, Australia, and New Zealand. Why can't we go back to these countries again?" For a carer in her 70s from the West Midlands (P1282), "We betrayed Australia and New Zealand when we became united with Europe." An insolvency practitioner in his 50s from the East Midlands (M1460) thought: "We should cultivate stronger ties with the Commonwealth with a view to future trading relationships." A teacher in her 60s from Scotland (W729) agreed: "I want trade agreements with the Commonwealth, USA, and other countries, instead of Brussels saying who we are allowed to trade with." This last quotation introduces two further parts of how Leave supporters frequently imagined Britain and its place in the world around the time of the referendum. Britain was imagined to be positioned, or to have the potential to be positioned, in a "wider world" beyond Europe. A market researcher in his 60s from the North West (P4287) wrote: "We should be very hopeful for the future, free to negotiate trade deals with the world … We can still have a healthy relationship with Europe, but also be open to the world." For some panellists, this world beyond Europe was populated by "growing" or "booming" economies. Here again is the woman in her 80s from the East of England (H260): "Most of us are optimists and look forward to facing the challenge of new trade links with economies that are booming in the world." And here is a carer in his 30s from the North East (N5744): I believe an economically tough period of adjustment after leaving the EU would indeed be inevitable, but in the long run I can't but feel that as a country we'd bounce back and may even emerge in a much stronger position by establishing new and better trade with growing economies like Brazil, India, and China.
For Leave supporters, Britain had the potential to be "free," "open," "stronger," "booming," but it was constrained by membership of the EU. This was another part of how Britain was imagined by many Leave supporters. The teacher in her 60s from Scotland (W729) put it like this: I have always had a 'wider world' mentality. I never felt we in Britain should first look to Europe to trade etc. I always felt that Britain was part of the wider world with regard to trade and so on. I feel that being in the EU has somehow confined Britainmaybe restricted her influence abroad, shackled Britain in some way, compared to pre-EU.
For this panellist, Britain was great in the past and could be great again in the future, if only it could break the confines or shackles of the EU. This view was repeated by many other Leave-supporting panellists. An electronics engineer in his 40s from the North West (L5642) wrote: "I hope that the UK can flourish without the EU restrictions. I am pretty sure this country can still be as great as what it was." A mechanic in his 60s from the East Midlands (T3155) offered the following analogy: "Leaving the EU will be, for the UK, similar to leaving a smoke-filled pub for fresh country air." He continued: "They have entangled the country into the gradual formation of a United States of Europe. We see the writing on the wall: get out, leave, disentangle before the final strangulation of 'Great Britain.'" In constructing and expressing their views around the time of the referendum, Leave supporters appear to have drawn heavily on images of Britain as an island separate from Europe and narratives of Britain as a once great military and imperial power that could be great again if released from the EU and freed to engage more in the wider world. We should add: these images and narratives echo those found in elite discourses by others (Grob-Fitzgibbon, 2016;Sykes, 2018;Wenzl, 2019;Whittaker, 2018). Indeed, the same can be said for another image promoted by elites in the period leading up to 2016: "island Britain" as a full containersmall, too close to the rest of Europe, and vulnerable to traffic of various kinds from France and other EU Member States (Cap, 2019;Sykes, 2018).
This imagewhich apparently contradicts the image of "island Britain" as great, separate from Europe, open to the worldwas used by many of the same panellists quoted in the paragraphs above. Recall from Section 2: popular imaginative geographies can be happily contradictory in a way not available to formal theories or political ideologies. Listen again to the insolvency practitioner in his 60s from the East Midlands (M1460): "The referendum has made people aware of just how many foreigners are here … These extra people are filling up the schools and hospitals … The country is full." And here again is the carer in her 70s from the West Midlands (P1282): "Our schools and hospitals are full to breaking point." Other panellists explicitly linked Britain's fullness to its island geography. For a retired journalist in his 70s from the North East (W633), Britain -"as an island … with finite space and resources"should worry about "the uncontrollable influx of refugees [to Europe] from North Africa and the Middle East." For a retired factory hand in his 70s from the East of England (C2579), "Our little island" is "fast becoming the dustbin of Europe"for "criminals" and "people who just want our benefits." Leave supporters, it would seem, imagined Britain to be an island separate from Europe, but also close to Europe; great and open to the world, but also small, full, and vulnerable to cross-border mobilities. These popular imaginative geographies echoed elite discourses regarding "island Britain" and were likely influenced by those discourses, though our study was not designed to confirm such a process of causation. Our original contribution in this section has been to demonstrate how Remain and Leave supporters imagined Britain around the time of the EU Referendum. Remain supporters constructed and expressed their opinions from narratives of post-imperial decline and images of a small, vulnerable Britain, at risk of being isolated in a dangerous world. By contrast, Leave supporters made sense of the issues by drawing on a different repertoire of narratives and images. There was the historical narrative of Great Britain, an island nation with a history of independence from Europe, of military and imperial greatness, of openness to the Commonwealth and the wider world, that became constrained by membership of the EU. There was the image of Britain as an island with finite resources, close to Europe and so vulnerable to various mobilities from Europe. And there was the image of a wider world where, unlike the rest of Europe, economies are booming and Britain could thrive. We now turn to another original contribution of the paper: a demonstration of how and why these and other popular imaginative geographies came to be used by citizens as they voted in the referendum. 5 | WHERE DOES VOTER DECISION-MAKING TAKE PLACE? Froud et al., (2016) have argued that national averages like Gross Domestic Product resonate little with citizens in a society characterised by uneven development and socio-economic polarisation. We see this in how MO panellists justified their vote for Remain or Leave. Few panellists mentioned such national averages. Instead, most wrote about their locality. Voter decision-making happened in local contexts. This was the case for Remain supporters like the retired social care manager in his 50s from Scotland (A3623), who wrote about how "The local council here got EU funding to rebuild and redesign our bin shelters and back courts," or the unemployed woman in her 20s from Wales (B5880), who wrote: I don't know a lot about politics/the EU, but I do know that I very definitely want to Remain. I know that my university, and by extension my university education, benefited greatly from EU funding … As a theatre practitioner, I know the loss of EU arts funding will be very difficult. And I know that parts of Wales are some of the poorest in Europe and again will be hard hit by the loss of EU funding.
Here, to justify her Remain vote, the respondent focuses not on "politics/the EU," which she admits to knowing little about, but on her own sector and region, which she knows to have benefited from EU funding. A similar focus on personal experience and the regional or local situation, which panellists felt could be known by themwhen much of the complexity of the referendum debate could not be known by themwas evident in the writing of Leave supporters.
Some panellists justified their Leave vote by writing of how their locality had changed for the worse because of immigration. "Most of the districts I knew and loved in London have changed beyond recognition," wrote a woman in her 80s (H260). She continued: "Foreign shops and restaurants are everywhere. Lots of people no longer speak English." A retired office worker in her 60s (C5991) wrote of Barking: "It feels like I am in a foreign country and it does not feel safe when you cannot understand what people are saying on the street." In the East of England, an unemployed man in his 40s (B5970) wrote: "It is literally possible to walk from one end of this town to the other and not hear a single word of English." In Scotland, a teacher in her 60s (W729) wrote: "You walk down a street, even up here, and all you hear are different languages spoken, or English spoken with a foreign accent." When asked to vote in the referendum, and faced with speculative claims and counter-claims about national averages like GDP, these Leave supporters took recourse in what they thought they could know: their own experience of the locality in which they live. They also took recourse in stories told by their family, friends, and neighbours: My sister-in-law, who has lived in London for 80 years, now tells me that 'slums' are returning. Migrants are living in sheds and garages because of the housing shortage … Her grandchildren are unable to get decent school places for their children. Her doctor's practice now has to employ translators on certain days to deal with immigrants who don't, or won't, speak English.
The above quotation is from a woman in her 80s from the East of England (H260). The following quotation is from a retired nursery teacher in her 60s from the West Midlands (M3408): My son has 30 years' experience as a chef and his wages have barely risen above living hand-to-mouth in that time, and he blames it on cheap labour from abroad. It leads to resentment. I have a number of friends who are teachers and they are literally drowning under the pressures of coping with up to 32 languages spoken in one school and the strain of dealing with such a diversity of culture and discipline issues. It leads to resentment.
These could be read as tales of an economy that fails to deliver decent wages and a welfare state struggling to deliver housing, education, and healthcare after years of austerity. Indeed, Warhurst (2016), among others, has criticised the Remain campaign for focusing not only on national averages, but also on what is threatened by Brexit (e.g., the national employment rate), as opposed to what is building resentment now (e.g., insecure and poorly paid jobs). Furthermore, the quotations in the previous paragraph could be taken as empirical support for Norris and Inglehart's (2019) theory of "cultural backlash" (against cosmopolitanism and multiculturalism), or Goodwin and Milazzo's (2017) argument that residents of areas undergoing fast ethnic change feel a "loss of control" (for more on Brexit, migration, and race, see Burrell et al., 2019). This is no doubt very important, and we broaden our discussion in Section 6, but our primary focus here is on how and where citizens arrived at their voting position. Panellist M3408 went on to write the following lines: The only facts are what we know and experience, and which affect our daily life. It's a fact to the parent who can't get their child into a local school. It's a fact if your child is one of a handful who can speak English in an over-subscribed class in a makeshift classroom. It's a fact if the doctor's appointment you really know you need, but are told you cannot have for another three weeks because the receptionist insists they are all booked.
The paragraph continues with "facts" about the local hospital, crematorium, and housing situation. Asked to vote on whether Britain should remain a member of the EU, this respondent, like many others, justified their vote in terms of what could be known confidently by them, which was their own perceptions of "daily life" in their own locality.
Voter decision-making took place in local contexts and utilised popular imaginative geographies of particular localities. But there was another location from which voter decision-making took place, also made clear by volunteer writing for MO. Following a campaign that most panellists found unhelpfulcharacterised by accusation and counter-accusation, fearmongering, misinformation, and liesvoter decision-making happened in the gut. This was the case for Remain supporters like this self-employed professional in his 40s from the South East (D4736): "No one really seems to have a clue … My gut feeling is to remain." It was the case for Leave supporters like the retired nursery nurse in her 60s from the West Midlands (M3408): "So many questions and no solid answers. What can we rely on then? … Gut feeling?" A carer in his 30s from the North East (N5744) explained why it was the case for so many panellists: The public is being bombarded from every angle with opinions from politicians about which decision is right or wrongthe only problem is that every MP offering an opinion seems to be scaremongering or appears to be equally vague on the facts as their rivals are. I believe people do have a civic duty to vote, though, so I will do so anyway, even if I have to trust my gut instinct. This panellist appears confused by the campaign. They don't really feel informed enough to cast a vote, whether for Leave or Remain. They do so anyway, because they feel a duty to votea feeling that is long-standing and widely held in British political culture (Clarke et al., 2018). To do so, in these circumstances, they rely on gut instinct.
What, though, do respondents appear to mean by "gut instinct" or "gut feeling?" Here, we are taken back to the literature and particularly to the discussion in Section 2 of popular imaginative geographies as fast-thinking heuristics. For this is what MO writers appear to mean by locating decision-making in the gut. Listen again to the retired nursery teacher in her 60s from the West Midlands (M3408): For every argument on one side, there is a counter-argument on the other. Are any of them based in facts? Well, apparently there are 'facts' to consider, but amazingly these can be totally different depending on which camp you belong to. Everything else is conjecture. Nobody can actually predict the outcomes. So how do we decide? Gut feeling, probably. Possibly just asking ourselves questions such as 'Do I want to be defined as British or European?', 'Do I want to retain our individuality and the freedom of our country to make its own mistakes and not have someone else, hundreds of miles away and a culture apart, make them for us?' She found the campaign unhelpful. Because it was a debate partly about the potential future consequences of a decision to Leave or Remain, it was characterised more by conjecture than facts. She therefore resorted to "gut feeling." What she appears to mean by gut feeling is the automatic, intuitive, emotional reasoning of System 1 or fast thinking. She appears to mean a form of political action reliant on cues and heuristics, including imaginative geographies made up of images like EU decision-making that happens "hundreds of miles away and a culture apart." To complete the point, now listen to three panellists who voted Remain: "I am not sure that … I am informed enough. My gut reaction is that we should stay in the EU. I feel that we work better together" (B5702); "My gut feeling is that it's better to be part of something and to try to influence changes from within than to sit outside of it" (T4715); "I have no real knowledge of whether Britain is better off in Europe … I just have a gut feeling that we should not be alone, that we should be allied to neighbouring countries whose ideals are generally more liberal … and similar to ours than many other countries in the world" (V3773). These respondents, like many others, did not feel informed enough to cast a vote. In relying on their gut, where decision-making was often located in writing for MO, they relied on imaginative geographies of Britain (potentially outside and alone), the EU (made up of neighbours who share Britain's liberalism), and the world (made up of other, dissimilar countries). In Section 4, we identified the popular imaginative geographies available to Leave and Remain supporters around the time of the EU Referendum. In the present section, we have shown how these imaginative geographies came to be used by voters as fast-thinking heuristics after a campaign that failed to provide citizens with trustworthy alternative resources.

| CONCLUSION
However the relationship between the UK and the EU develops in the coming years, debates about what voters knew, thought, and felt around the time of the 2016 referendum will continue for many years, if not decadesas they have done for other especially important moments in British political history (Moss et al., 2016). In this paper, we have made original contributions to these debates in the following ways. We have established popular understandings of Brexit, Britain, the EU, and the world, not by inferring them from the characteristics of citizens, or the characteristics of the regions/localities in which citizens reside, or elite discourses, but by listening to what citizens say when given the opportunity to write in their own terms. Methodologically, we have brought a new dataset and associated analytical approach to these debates: volunteer writing for MO. Conceptually, we have read these data for popular imaginative geographies, having developed that concept from literatures in Geography and Political Scienceso that popular imaginative geographies are understood as fast-thinking heuristics used by citizens to make sense of issues and act politically when invited to do so in difficult circumstances. Our main original contribution, however, has been empirical.
We have identified the range and content of imaginative geographies available to British citizens around the time of the EU Referendum. Many Remain supporters imagined a small, post-imperial, vulnerable Britain, under threat of isolation from Europe and exposure to a chaotic, dangerous, uncertain world. Many Leave supporters imagined a Janus-faced "island Britain" that on the one hand was separate from Europe ("island Britain" as a once great military and imperial power that could be great again if released from the EU and freed to engage more in the wider world) and on the other hand was a full container ("island Britain" as small, close to the rest of Europe, and vulnerable to mobilities across Europe's borders). These different imaginative geographies shared a similar starting point, despite their clearly opposed end points. 1 They began from a sense of loss, with Britain imagined as once great and powerful, but now post-imperial. For many Remain supporters, a country left vulnerable by long-term and perhaps irreversible declinethought to pre-date Britain's entry into the Common Marketneeded protection from within the EU. For many Leave supporters, a country in decline only recentlysince joining the Common Market in 1973, or the Maastricht Treaty of 1992, or EU enlargement in the 2000sneeded freedom from the EU, after which greatness would be attainable once again.
Having demonstrated the widespread existence of these imaginative geographies among groups of MO panellists (Leave and Remain supporters), and argued that categories and storylines shared widely by MO panellists might be taken as indications of the existence of a set of cultural resources in broader society (on which the panellists were drawing to make sense of the issues, very plausibly alongside other citizens), we also demonstrated how imaginative geographies came to be used in the act of forming a position and voting in relation to the referendum question. Referendums are particular democratic events that ask a great deal of votersoften in the form of questions about complex issues, the answers to which cut across traditional party linesleaving voters with needs for cues and other shortcuts to help them act. In 2016, many voters found the EU Referendum campaign unhelpful. It focused on national averages perceived to be meaningless in a society characterised by uneven development. It focused on what might happen in the future and was perceived to be characterised more by speculation than fact. Every claim appeared to be answered with a counter-claim, leaving voters adrift in a sea of distrust. The Remain campaign was especially unhelpful to voters, failing to circulate positive images and narratives of Britain in the EU, and failing to address the resentments felt by many British citizens in 2016. In this context, feeling a duty to vote, many citizens became dependent on two things: local knowledge (their own perceptions regarding everyday life in their own localities) and gut feeling. In this context, voters became dependent on fast-thinking heuristics, with popular imaginative geographies to the fore.
These contributions, focused on the content of popular geographical imaginations and how they were used by voters, help us to understand the outcome of the referendum. They add complementary empirical support to existing theories based on inferences from demographic or economic data at the individual, local, or regional level. First, they add evidence of what people wrote when given the opportunity to do so in their own terms. Here, we see that many Leave supporters did write nostalgically about imperial Britain (as predicted by Dorling & Tomlinson, 2019), or did write fearfully or angrily about ethnic and cultural change (as predicted by Goodwin & Milazzo, 2017;Norris & Inglehart, 2019). Second, these contributions add evidence that fleshes out the bones of existing theories. We see in the MO diaries what connected apparently separate theories. It was the figure of "island Britain" that allowed Leave supporters to imagine Britain as great in the past (a sea power, separate from Europe, with its own empire), vulnerable in the present (to free movement of immigrants from Europe, threatening to fill up this little island), and potentially great in the future (a trading nation, free from EU regulations, open to the world).
Finally, these contributions add to understandings of why people voted Remain. This question has been relatively neglected by existing studies, perhaps because the focus has been on explaining the referendum result (the victory of Leave), but perhaps also because of an assumption that Remain was the rational position, the position supported by most scholarsthe position, therefore, needing least explanation. To this question, we can make two responses based on our analysis. First, many Remainers justified their vote in the referendum by referencing Britain's small size, vulnerability, and potential isolation in a world increasingly characterised by danger and uncertainty. Second, in a context of claim and counter-claim, misinformation, and perceived fearmongering, Remain positions were no less based on popular imaginative geographieson fast-thinking heuristicsthan Leave positions. They were no less based on gut feeling or instinct, as opposed to facts about consequences.
Where do these contributions leave our understanding of the referendum result? A full understanding requires connecting up these contributions to other bodies of research. We need to know the content of popular imaginative geographies and how they were used by voters, but also the distribution of imaginative geographies between groups of votersand not only Leave and Remain voters, but also more disaggregated social groups. This would allow for a better understanding of where these imaginative geographies came from, with a particular focus on the material circumstances and associated experiences of particular social groups. It would require an exchange between more intensive studies of popular imaginative geographies and more extensive studies of public opinion, voter demographics, and the characteristics of localities/regions. A better understanding of the origins of these imaginative geographies also requires more research on the relationship between elite and popular discourses of Brexit. It is likely that popular discourses are framed by the elite discourses of politicians and journalists. More research is needed, though, to establish that relationship empirically. Only then will a full understanding of Brexit and popular imaginative geographies be within reach.

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of the article.