National Identities and the 2014 Independence Referendum in Scotland

Introduction

1.1 On the 18^th of September 2014 a referendum took place to determine whether Scotland should remain part of the United Kingdom or become independent and end a political union that had endured for more than 300 years. This was not the first referendum to decide Scotland's constitutional future. In 1979 a proposed Assembly to give Scotland more formal political autonomy within the UK failed to achieve the necessary support from the electorate. In 1997 a similar initiative for a 'Parliament' with more substantial autonomy than the proposed 1979 Assembly easily achieved the required popular mandate, and the new Parliament opened in 1999. But the 2014 referendum was unprecedented in the UK in that it concerned the complete political secession of one of its constituent nations. It was also a rare example, even at a global level, of a democratic popular vote to determine the possible secession of part of an existing state to form a new independent state.

1.2 The referendum campaign, the vote itself and its aftermath stimulated a huge amount of debate and coverage, and this has been re-enlivened by the result of the 2015 UK General Election in Scotland, with the Scottish National Party winning 56 of 59 seats. Although, not surprisingly, much of the discussion of the referendum and its aftermath has been focused on the political dimensions, a number of sociological perspectives were offered both prior to and after the vote. For example, prospectively, Bechoffer and McCrone (2014) discussed the status of British identity in the run up to the referendum, while Rosie (2014a) examined claims about referendum voting intentions and the potential impact of independence on religious groups in Scotland. Retrospectively, the British Sociological Association's Discover Society offered a rapid response to the referendum outcome in several contributions concerning issues such as democratic engagement and the future of nationalism in Scotland (Rosie 2014b) and the significance of gender in the campaign and vote (Rummery 2014).

1.3 This paper has a different aim: to address the ways in which the referendum may be understood through reference to national identities and belonging. I do this firstly by reflecting briefly on who was (and was not) allowed to vote. The determination of the referendum franchise was not merely a matter of practical necessity: to some extent it reflects a disjuncture between political and sociological perspectives (and evidence) regarding the nature of national belonging in Scotland. I then develop this examination of the sociological and the political dimensions of nationhood by bringing together recent and historical survey data on people's subjective national identities and political attitudes. This analysis generates some unprecedented findings, suggesting the potential for more highly politicised national identities in Scotland. I then conclude by reflecting on how the concept of 'social union', which so far has been only quite weakly articulated by politicians and seldom explicitly researched by social scientists, might help us make sense of some of these findings.

National Belonging and the Referendum Franchise in Scotland

2.1 Renan famously described the very existence of a nation as 'an everyday plebiscite' (1994 [1882]: 17), but in a less figurative and more literal sense an actual plebiscite in which the people of a nation vote both reflects the existence of that nation and something of its nature. Tierney states that '… the constitutional referendum by definition implicates an anterior act of demotic border-drawing — the framing of the collective self who will perform an act of constitutional self-determination and in doing so explicitly articulate itself as a constitutional people' (2012: 58-9). While entitlement to vote is inevitably a factor in any election, in a referendum concerned with the very future of the state itself (or part thereof) the precise nature of 'the people' is potentially both a more vexed and sociologically interesting matter, which also takes on added complexity in contexts of sub-state nationalism such as that found in Scotland, where state and national demos obviously differ. Tierney observes that '… the constitutional referendum sets the boundary of the people by way of both territorial demarcation and franchise rules' (ibid.: 59). Both these dimensions were observable in the case of the referendum on Scottish independence. The profound impact that a Yes vote would have had on the UK as a whole raised the question of whether 'the people' of the UK should all have an equal say through voting rights in the referendum, but this possibility was not ever seriously considered and the territorial demarcation of the people was based on the borders of Scotland, which are long settled and uncontroversial.

2.2 Potentially more controversial was the second dimension concerning the nature of the franchise within these borders. As Tierney (2012: 59-60) observes of constitutional referendums more generally, the two most problematic questions tend to be: 'are there or should there be people among those resident within the territory not entitled to vote? And, are there people resident beyond the territory to whom, nonetheless, the franchise is warranted?'. These are also rather more sociologically relevant and interesting questions because they go to the heart of how national 'belonging' or 'membership' may be determined or defined^[1].

2.3 While a wealth of sociological evidence on popular opinion about the key 'markers' of Scottishness (those personal characteristics that people consider most important in determining who may belong to the nation) indicates that birthplace is clearly the most important factor (Bond 2006;Kiely et al. 2001;Kiely et al. 2005a;McCrone & Bechhofer 2008), this was not reflected in the franchise for the 2014 referendum in Scotland. Those entitled to vote were essentially the same as for local government and Scottish parliamentary elections, and indeed for the 1997 referendum on devolution. An important exception was that 16 and 17 year olds were included, which was itself an unprecedented feature for any kind of major political vote in a UK context. All UK citizens normally resident in Scotland were entitled to vote, but so too were citizens of other EU member states and 'qualifying citizens' of Commonwealth countries^[2].

2.4 So, although a large majority of those who could vote were born in Scotland, for the purposes of the referendum franchise the key marker of Scottish national identity – birthplace – was effectively irrelevant, as those resident in Scotland but born in other parts of the UK, EU and Commonwealth had an equal say in determining Scotland's constitutional future^[3]. This principle extended to those born in Scotland but living in other parts of the UK: in England alone there are approximately 700,000 Scottish-born residents (Office for National Statistics 2011) who were unable to vote in the referendum, as were Scottish-born people living in other parts of the world and those not born in Scotland who might nevertheless consider themselves Scottish on the basis of parentage or more distant ancestry. For example, in the USA in 2009 nearly 6 million people reported having Scottish ancestry (US Census Bureau 2012)^[4]and in Canada in 2006 nearly 5 million (4,719,850) did so (Statistics Canada 2006)^[5]. In Australia, the 2006 Census reports 1.5 million people claiming Scottish ancestry (see http://www.abs.gov.au/ausstats/abs@.nsf/7d12b0f6763c78caca257061001cc588/5a47791aa683b719ca257306000d536c!OpenDocument). Although the exclusion in particular of the many Scottish-born people living in England did generate a degree of grievance, there was no serious and concerted political agenda on the part of any of the interested parties to extend the franchise beyond Scotland in any way. Although not extended to all (i.e. non-EU, non-Commonwealth citizens), the application of the marker of residence to determine the franchise was therefore relatively uncontroversial, despite research evidence indicating that this is popularly considered to be a relatively weak marker that may be put forward as the basis for a Scottish identity. Claiming to 'belong' to Scotland based largely on residing in Scotland – even when that residence might be enduring and coupled with a firm commitment to remain in Scotland – is often seen as difficult both by migrants themselves and those who are born in Scotland (Bond 2006;Kiely et al. 2001;Kiely et al. 2005a;McCrone & Bechhofer 2008;McCrone & Bechhofer 2010).

2.5 Undoubtedly, this approach was to a considerable extent pragmatic. It made the determination of eligibility relatively straightforward in ruling out people who would have been more difficult to evaluate (e.g. proving Scottish birthplace and/or ancestry); it meant that the electorate to be addressed by both campaigns (supporting and opposing independence) was fairly clear and territorially-focused; and it was based on the established precedent of other elections in Scotland (excluding UK-wide Westminster elections) and the 1997 referendum. But the referendum franchise might also be viewed as having wider sociological significance. As Tierney (2012: 60) argues, 'the way in which the franchise issue is dealt with in the design of a referendum can tell us much about the type of nationalist ideology dominant in a specific state or territory; for example, particular rules of inclusion or exclusion reveal whether the vision of the nation that prevails is more or less 'civic' or 'ethnic' in orientation'. To an extent, then, the franchise for the referendum on Scottish independence reflects a conception of nationalism in Scotland as largely liberal rather than essentialist and exclusive (see e.g.Hearn 2014).

2.6 However, it is important to recognise that the simple distinction between civic and ethnic nationalism has been widely critiqued (see e.g.Brown 1999;Brubaker 2004;Yack 1999). This has entailed a questioning of whether such ideal types may be straightforwardly applied to specific cases, and the extent to which it is legitimate to describe civic nationalism as fundamentally liberal and inclusive and ethnic nationalism as illiberal and exclusive. While the referendum franchise certainly accorded with an elite consensus in Scotland that views Scottishness as largely inclusive and open and encourages 'new Scots' who may claim this identity on the basis of neither birthplace nor ancestry, this begs the question of what it is that such people would be identifying with. As Brubaker (2004: 138) observes, even putative examples of civic nationalism such as that found in Scotland are based on certain characteristics (common values, practices, histories) that one might associate with more ethnocultural conceptions of nationalism (see also Yack 1999: 106-7). Further, not only might voluntaristic national identification come with certain conditions attached, this voluntarism is itself limited by wider factors: even the acquisition of Scottish residency is clearly not equally available to all peoples of the world. Finally, it is not necessarily the case that elite perspectives accord with popular conceptions of nationhood in Scotland, which may be rather more exclusive (Bond 2006).

National Identities and the Referendum Vote

3.1 If the referendum franchise might to some extent be understood sociologically in relation to questions about national identity, then so too can the referendum vote. In this paper I do not consider direct evidence concerning people's reports of how they actually voted, partly because such data are not widely available at the time of writing. But I do consider a substantial and well-established body of evidence, directly comparable over a considerable time period, which concerns the perspectives of the people of Scotland on the constitutional question and related political issues. Importantly, I compare this evidence to similar data collected in the period immediately before the referendum (May-August 2014), which suggests some unusual, interesting and potentially important developments. In particular I relate political perspectives to people's own subjectively expressed national identities. Because at its heart the referendum debate concerned whether and in what ways Scotland should remain part of Britain, I use an established measure of subjective identity that explicitly invites people in Scotland to express the relative priority of Scottish and British identities. This is the so-called 'Moreno' question^[6] based on a 5-point scale with the following options: Scottish not British; More Scottish than British; Equally Scottish and British; More British than Scottish; and British not Scottish.

3.2 There is a well-established body of research examining how subjective identities on this scale may be associated with key political variables, most prominently constitutional preference (essentially, independence, devolution, or no Scottish Parliament) and which political party people identify with most closely (see e.g. Bond 2000;Bond & Rosie 2002;Bond 2009;Bond & Rosie 2010;Rosie & Bond 2003). Although such associations in Scotland have sometimes been described as 'weak' (see e.g. McCrone & Bechhofer 2008: 1263), this is probably overstating the case somewhat given that there are clear associations to the extent that people who identify as exclusively Scottish or at least prioritise Scottishness are much more likely to support independence and/or the Scottish National Party (SNP) than are people who feel themselves to be equally Scottish and British, or have a prioritised or exclusive British identity. But, as we will see, it is also certainly true that there has indeed been a clear 'non-alignment' (Bond 2000) between subjective national identities and these key political variables. As Hearn states '… in the Scottish context, correlations between national identity, party support, and constitutional preferences are loose and shifting' (2014: 507).

3.3 But re-assessing these correlations using evidence from the referendum period generates some interesting results. It suggests not merely a continuation of established patterns, but that either the referendum may have brought about a particular, although perhaps temporary, set of conditions in which subjective national identities became much more closely 'aligned' with key political perspectives, or alternatively that it may have instigated a more enduring alignment between identities and political positions.

3.4 As a useful starting point in the analysis, table 1 shows trends in national identities using the 5-point Scottish-British 'Moreno' scale over a period of more than 20 years. These trend data are somewhat counterintuitive given what we might expect about the effect of a newly established and prominent political institution on subjective identities. Specifically, rather than a shift towards more strongly prioritised Scottish as opposed to British identity following the re-establishment of the Scottish Parliament, the evidence suggests the opposite. This is most noticeable in the trend decline of those in the Scottish not British and More Scottish than British categories, from a high point of more than 2/3 in the year that the Scottish Parliament was re-established (1999), to a (narrow) minority in the year of the independence referendum. A further point to note about the data from 2014 is that this marks the first time that the equally balanced dual identity was the modal category^[7], and indeed the only instance shown, with the exception of 2001, where the More Scottish category is not the most popular. Hence there is no evidence here to suggest a decline in people's self-identification as British as the Scottish Parliament has become established, developed and taken on new powers. This is particularly notable given that, since 2007, the parliament has been led by a governing party (the SNP) whose raison d'etre is the dissolution of the (British) political Union, and which is currently enjoying unprecedented success in electoral performance and recruitment of new members. This immediately suggests that there is no simple correspondence between people's subjective identification as Scottish and/or British and political behaviour, attitudes and structures in Scotland: national identities are socially as well as politically shaped.

3.5 We should also note that in some respects national identities in Scotland have not changed so much over recent history – most notably the proportion prioritising Britishness, although it has increased very marginally in recent years, remains a very small minority when compared to those who prioritise Scottish identity or indeed see themselves as exclusively Scottish.

Table 1. Trends in subjective national identities in Scotland, 1992-2014

Sources: Data from 1999-2014 are from the Scottish Social Attitudes surveys (ScotCen Social Research 2014); data from 1992 and 1997 are from Scottish Election Surveys (seeBond & Rosie 2002). All data excluding sample sizes have been weighted. All percentages have been rounded to the nearest whole %.

Key: SnB (Scottish not British); S>B (More Scottish than British); S = B (Equally Scottish and British); B>S (More British than Scottish); BnS (British not Scottish); Oth (Other, None or Don't Know).

3.6 We now turn to the question of the 'alignment' of these expressed identities with the key political variables of constitutional preference and party identification. We initially focus only on how support for independence relates to national identities, shown in Table 2. In this and the remaining tables below, only selected years are shown from all the Scottish Social Attitudes surveys over the period. For tables 2 and 4 I have simply presented data at five year intervals for the decade immediately following the re-establishment of the Scottish Parliament in 1999, and increased the frequency to two year intervals for 2010-2014. These data give a general sense of the associations shown in the tables and their development over time. Although there are inevitably some fluctuations in the data not included, no obvious 'outlier' years have been omitted. The aim is to give a reasonably faithful representation of the data over the period while being parsimonious in order to maintain clarity and avoid the presentation of an excessive amount of information.

Table 2. % of each national identity category supporting Independence, selected years 2000-2014

Sources: Scottish Social Attitudes surveys. All data excluding sample sizes have been weighted. All percentages have been rounded to the nearest whole %. The More British than Scottish and British not Scottish categories have been combined and all those not answering or who responded 'don't know' to the question on constitutional preference have been excluded.

3.7 Overall, Table 2 shows a clear association between subjective national identity and support for independence. In each year those who describe themselves as exclusively Scottish are clearly the most likely to support independence and those who describe their identity as more Scottish than British are in turn much more likely to support independence than are those who do not prioritise their Scottishness. In all years but one those with an equal dual identity are more supportive of independence than those who prioritise their Britishness, but the differences between these groups are much less marked.

3.8 However, there is also clear evidence to suggest that, especially in the period before 2014, the association between identities and support for independence is not especially strong. Although there is evident fluctuation in the data, only around half of those who describe themselves as exclusively Scottish support independence (this is also true of other years not shown in the table). We can also see that clear minorities – approximately between one-quarter and one-third – of those in the More Scottish category also support independence. These data do not suggest that a prioritised Scottish identity is necessarily politicised, at least with regard to a desired correspondence between subjective national belonging and political representation in the form of independent statehood. This further substantiates the earlier point made in relation to Table 1: key political-constitutional attitudes do not map on neatly to subjective national identities in Scotland, suggesting the latter have an important social dimension. Equally, Table 2 shows that support for independence among those who see their Britishness as equally or more important than Scottishness, or indeed do not see themselves as Scottish at all, is perhaps higher than we might anticipate. Typically, small but far from trivial minorities in these identity categories express support for an independent Scotland and we must assume that this support is largely based on factors other than their self-conceived sense of national belonging.

3.9 Despite an obvious pattern of association, then, 'non-alignment' is also clearly evident in the pre-referendum period. On the one hand, we might say that it is somewhat counterintuitive that so many who profess to feel no sense of Britishness at all seem content to continue with a system of government in which the British dimension seems paramount. But on the other hand we might say that at least a partial explanation for this lies with the existence of a devolved Scottish Parliament within the broader context of the UK state: a strongly prioritised Scottish as opposed to British identity might be represented politically by a level of national autonomy short of full statehood. This is explored in Table 3 below, which presents a fuller analysis of the association between national identities and constitutional preferences. In the interests of economy and space, only two years are shown: 2014 and 2012. This highlights a contrast (which will be discussed further below) between the period immediately before the referendum (2014) and the earlier period represented by 2012. Although not shown in the table, the figures for 2000 and 2010 are much more similar to 2012 than they are to 2014. This is not true to the same extent for 2005, mainly because support for independence overall was higher in that survey than in the others years pre-2014, but, as already shown in Table 2, this support among the Scottish not British group was still substantially lower than in 2014.

Table 3. Preference for the future government of Scotland by national identity category, 2014 and 2012

Sources: Scottish Social Attitudes surveys. All data excluding sample sizes have been weighted. All percentages have been rounded to the nearest whole %. The More British than Scottish and British not Scottish categories have been combined and all those not answering or who responded 'don't know' to the question on constitutional preference have been excluded.

3.10 Pre-2014 trends in support for a devolved Scottish Parliament by national identity to some extent indicate a degree of consensus as opposed to 'alignment'. This is consistently the most popular option among all groups who express any sense of Britishness at all, but large minorities (ranging from about one-third to just less than half) of the exclusively Scottish group also typically prefer the current devolved settlement to full independence. In fact as recently as 2010 devolution was actually more popular among those expressing a Scottish not British identity than was independence.

3.11 However, as we now know, when constitutional preferences were formally put to the test in the 2014 referendum, the outcome, although rejecting independence, did not reflect a consensus around devolution to the extent that the pre-2014 survey data might suggest, with 45% voting for independence. An explanation for this may be found if we focus on support for independence by national identity in 2014, shown in both tables 2 and 3. Although these data were collected before the referendum, during the period of the fieldwork the campaign was already intense and the vote itself less than four months away. These data show a much stronger alignment between identities and support for independence compared to previous years, at least if we focus on the Scottish not British and More Scottish categories of respondent. Support for independence accounted for more than two-thirds of the exclusively Scottish group compared to typically less than half in previous years and there was a 40 percentage point difference between support for independence and devolution in the Scottish not British group in 2014, which was quite unprecedented. Even in the 2005 survey, when, as noted above, support for independence across all national identity categories was relatively high, this percentage point difference was only 24, and more commonly in the pre-2014 surveys levels of support for independence and devolution are quite similar among the Scottish not British group. Further, among the More Scottish than British respondents support for independence in 2014 increased from a figure typically between one-quarter and one-third to nearly half of such people. In the context of the wealth of trend data on '(non-) alignment' that we have from previous years these findings are clearly anomalous and interesting. It seems that in the context of the referendum campaign more people's identities were politicised to the extent that there developed a much stronger association between prioritised Scottishness and support for independence in particular.

3.12 Before we seek to explore this phenomenon further by attempting to establish any other possible evidence of closer alignment between subjective national identities and political attitudes, it is however important to add two qualifications. First, there is no evidence of such alignment among those at the opposite, most British end of the identity scale. If stronger alignment were taking place at all points along this scale, we would expect that only a tiny minority of the most strongly British identifiers would support independence, but in fact the proportion doing so in 2014 was the highest since 2005. Second, and perhaps most importantly, until we gather similar data from subsequent years we cannot establish whether such alignment as does appear to have taken place was solely related to the context of the referendum campaign itself, or whether the referendum and its aftermath will indeed mark the beginning of a new trend in which prioritised Scottish identity in particular is much more likely to be politicised with respect to a desire that this identity is reflected in full independent statehood.

3.13 Alongside views on Scotland's constitutional status, we can also examine how identities relate to individual political perspectives by examining data on party identification. This is preferred to analysing people's voting behaviour (recalled or intended) because it avoids the possible distortions of tactical voting and instead assesses which, if any, political party people identify with most strongly in a general sense. In Scotland, the extent to which we can say that any party is necessarily 'more Scottish' than others in any kind of essential way is a complex matter. A particular dimension of politics in Scotland is that all the major parties describe themselves as 'Scottish' and, as Hearn (2014: 508) points out, '… in Scotland, normal politics is not a struggle between nationalism (represented by the SNP) and something else; it is between contending nationalisms'. However, we may speak of 'alignment' in the sense of how these parties might be associated with different constitutional perspectives. Here we focus only on those four parties that have consistently attracted the highest levels of support in Scotland over the period for which data are analysed, i.e. the SNP, Labour, Liberal Democrats and Conservatives, although an interesting development in more recent political history has been the explicit support for independence voiced by some of the smaller parties, most obviously the Greens and various socialist parties.

3.14 Given their name and political raison detre, we would expect that if identities and political preferences are indeed aligned, then a large majority of SNP supporters would view themselves as exclusively Scottish or at the very least more Scottish than British. Labour and the Liberal Democrats are part of wider UK party structures and neither supports independence. Alongside the Conservatives, they campaigned (under 'Better Together') against Scottish independence in the referendum, but both had key roles in the Constitutional Convention that laid the groundwork for Scottish devolution in the 1990s and it was the (UK) Labour party that passed legislation for the 1997 referendum leading to the re-establishment of the Scottish Parliament in 1999. These are therefore parties among whose supporters we might expect dual Scottish/British identities to feature particularly strongly. Finally, the Conservatives have now endured a long period of comparative political failure in Scotland, and this failure is regarded by some as at least partly attributable to popular perceptions that they are essentially an 'English' party – i.e. related to a perceived lack of 'Scottishness' (Bednarek 2011;Hopkin & Bradbury 2006;McCrone 2001: 114). They are also, most obviously among all the parties, the most strongly associated with the political Union of Britain: in Scotland their official name is the Scottish Conservative and Unionist Party, and this strong unionism was reflected in their consistent opposition to devolution before the 1997 referendum. Overall then, if national identities and party political identification are aligned then we would expect Conservatives to have the most British subjective national identities.

3.15 Table 4 shows trends of the national identities of those who identify with each of the four main parties, using the same years and identity categories as were examined with regard to support for independence in Table 2. The data exhibit some similarities to the relationship between identities and constitutional preference in that: a) there is a clear association and the nature of that association is broadly as we would expect; b) there is also considerable evidence of 'non-alignment'; c) the most recent 2014 data from the referendum period shows some interesting differences to established trends. Points a) and b) have been well-established in previous research (Bond 2000;Bond & Rosie 2002;Bond 2009;Bond & Rosie 2010;Rosie & Bond 2003), and the table shows the consistency of these patterns over time, albeit with some fluctuations.

Table 4. National identities of those who identify with each major political party, selected years 2000-2014

Sources: Scottish Social Attitudes surveys. All data excluding sample sizes have been weighted. All percentages have been rounded to the nearest whole %. The More British than Scottish and British not Scottish categories have been combined. Those who professed another national identity or none are not represented in the table, hence percentages do not sum to 100.

3.16 Taking SNP identifiers first, the data are broadly as we would expect to the extent that very few prioritise a British over a Scottish identity and typically quite small minorities express an equal dual identity. A Scottish only identity is also the most common among this group in all surveys but one. However, there is evidence of non-alignment between national and party identification in that in all but one survey a majority of SNP identifiers see themselves as British to some extent. With regard to whether the 2014 data represent a departure from established trends for SNP identifiers, this does not appear to be true because the distribution of identities seems to be fairly typical. Although there was not much scope for SNP identifiers to become 'less British', we might have expected that if national and party political identification did become more aligned in the context of the referendum, then a much larger proportion of SNP identifiers would express exclusively Scottish national identities, but this is not the case.

3.17 With regard to Labour and Liberal Democrat identifiers, as expected dual identities are very prominent and it is difficult to speak of any 'non-alignment' here given that we would not have anticipated that identifiers with these parties would cluster towards either extreme of the Scottish-British scale. Patterns of national identity among these groups largely reflect those for people in Scotland as a whole, in that while dual identities are expressed by a large majority, it is clear that Scottish rather than British identity is most likely to be prioritised. But unlike SNP identifiers there is some evidence for these two parties that 2014 was different to previous surveys. This is most evident in the low proportions prioritising their Scottishness and high proportions expressing an equal dual identity compared to previous surveys. Among Labour identifiers, 2014 is the only year shown where only a minority (albeit nearly half) prioritised their Scottishness, and the only instance where an equal dual identity was the most commonly expressed. A similar phenomenon is observable among Liberal Democrat identifiers: in 2014 a much lower proportion than in previous surveys prioritised a Scottish identity and a (much) higher proportion saw themselves as equally Scottish and British.^[8] Although the data might also indicate that these patterns are a continuation of pre-established trends,^[9] nevertheless the extent of Britishness among those who identify most closely with these parties was especially marked in 2014. Therefore it is possible that the context of the referendum encouraged the expression of an equal Scottish-British identity among supporters of these two parties, both of which stressed the benefits of the Union ('better together') during the referendum campaign.

3.18 Turning finally to Conservative identifiers, they perhaps provide the most interesting illustration of a shift from non-alignment to alignment with national identities. In earlier surveys in particular, the profile of Conservative supporters is rather more Scottish than we might expect - not markedly different from Labour and the Liberal Democrats and with those prioritising Scottish identity clearly outweighing those prioritising Britishness. Although more recent surveys tend to show a trend decline in the proportion prioritising Scottishness among this group, the data for 2014 are particularly marked in this respect with only a small minority prioritising Scottishness, to the extent that this is the only survey in which Conservative supporters have a more British than Scottish profile. It is also the only survey in the table in which a majority of Conservative identifiers have an equal dual identity. This pattern is more in line with what we might expect given the nature of the Conservative party in Scotland and the UK, but it does not reflect previous data and trends: this appears to be another example where the specific context of the referendum has encouraged an alignment of political attitudes and national identities.

3.19 This alignment might be assessed further by bringing together both dimensions of the political attitudes that have been examined (constitutional preference and party identification) and examining how these articulate with national identities. This approach has again been employed in previous research (Bond 2000;Bond & Rosie 2002;Rosie & Bond 2003). Focusing only on those who express an exclusive or prioritised Scottish identity in relation to Britishness – i.e. those who describe themselves as Scottish not British or More British than Scottish – we can examine the proportions who a) both support Scottish independence and identify with the SNP; or b) support neither Scottish independence nor identify with the SNP.

3.20 The findings substantiate our earlier evidence concerning the unprecedented extent of alignment between national identities and key political variables in the context of the 2014 independence referendum. Among those who see themselves as exclusively Scottish, in the surveys prior to 2014 only quite small minorities (typically between one-fifth and one-quarter) both supported independence and identified with the SNP. At the same time, pre-2014, typically not much less than half the respondents in the Scottish not British group in fact adopted neither of these political positions. The pattern in 2014 presented a clear contrast to previous data, in that for nearly half of this group their exclusively Scottish national identity was fully 'aligned' with both 'nationalist' political positions, and only a quarter of this group adopted neither of these positions. Among the More Scottish than British group, not surprisingly the extent of non-alignment we observed among the exclusively Scottish respondents is even more marked in the surveys before 2014: typically only 10-15% support both independence and the SNP, and a clear majority (usually around 60%) support neither. But again alignment is markedly closer in 2014: the proportion aligned with both political positions increased to nearly a third, and only a minority (42%) supported neither independence nor identified with the SNP.

3.21 Of course, even in 2014, these data still do not suggest a very strong association between exclusive and prioritised Scottish national identities and politically nationalist attitudes. This is most obvious with regard to the Scottish not British group, for whom a sizeable minority continue to disassociate themselves from what we would expect to be the most obvious political manifestations of this identity. But the data do confirm that the referendum may have led either to an unusually close but perhaps temporary alignment of national identities and political opinion, or maybe the beginning of a period of closer alignment.

3.22 Thus, although we have observed that national identities in Scotland (Scottish and/or British) need to be thought of as social as well as political phenomena, it is possible that these identities are more politicised than ever before in the contemporary context, or at least that there is the potential for them to become more politicised when key decisions are at stake regarding Scotland's status as a 'stateless' or 'understated' nation (McCrone 2005), or a formally independent state. As Hearn (2014: 511) observes, 'nationalism […] is normal. It becomes peculiarly visible when aligned with options of territorial secession, as in the Scottish case, with the possibility of separation between Scotland and the rest of the UK'. This is significant given the unexpectedly close outcome of the referendum vote and its political aftermath, with intense debate about the nature of Scotland's autonomy and a huge surge in both electoral support for and membership of the key party for which the referendum result might have been a crushing defeat – the SNP. All these developments suggest that the debate about Scotland's political future is likely to be active rather than dormant in the years ahead.

Discussion and Conclusion

4.1 The argument and evidence in this paper highlights the complex relationship between national belonging in a social and political sense. Many of those who contributed to the decision on Scotland's constitutional future in 2014 would not understand themselves as subjectively Scottish at all, and yet because of the nature of the referendum franchise, insofar as we may speak of 'citizens' of a nation that is not also an independent state, they were and continue to be, in Marshall's (1950) terms, 'political citizens' of Scotland. Equally, the political rights given to those deemed to 'belong' on the basis of residence were not extended to include non-residents who might have felt a greater sense of national belonging on the basis of markers of national identity that most research has shown to be more salient in a Scottish context than is residence – namely, ancestry and (especially) birthplace (Bond 2006;Kiely et al. 2001;Kiely et al. 2005a;McCrone & Bechhofer 2008). This was especially evident for those born in Scotland (many of whom will also have been born of Scottish parents) but resident in other parts of Britain, who did not have the capacity to contribute to the determination of Scotland's future in the 2014 referendum.

4.2 Migration across national borders within the same (UK) state means that for many people in the UK, their national origins and subjective sense of national identity does not necessarily map on to the specific sub-state nation in which they reside. One consequence of this is the cross-border networks of family and friendship which represent one part of the broader concept of 'social union'. This term appears to have been in political currency since at least the late 1990s (see Tickell 2012) and is most closely associated with the SNP. The party has used the concept as a means to emphasise that the end of political union that would be brought about by Scottish independence would be paralleled by the continuation of various other bonds and affinities between the different UK nations (see e.g. Salmond 2012). As well as family and friendship, the other most commonly cited dimensions of social union relate to shared history and language; economic trade; cross-border labour mobility; and shared cultural reference points (e.g. in television and sport) (see Salmond ibid.;Scottish Government 2013; Tickell ibid.)¹⁰.

4.3 Although 'social union' has not been explicitly addressed by social scientists, it has been researched more implicitly to some extent with regard to the different dimensions of Britishness that it ostensibly entails, and how people's subjective national identities relate to these dimensions. Many of these studies are based on qualitative evidence, especially drawn from in-depth interviews. For example, Hopkins et al. (2006) and Bond et al. (2010) have explored how labour and educational mobility within Britain might be related to national identity. A large body of work by Bechhofer, Kiely, McCrone and colleagues (e.g. Kiely et al. 2005b) and Condor, Abell and colleagues (e.g.Condor and Abell 2006) has examined people's identification as British (and as Scottish or English), and the cited examples of their work show how elements of shared history inform people's national identification (or indeed non-identification). It is notable that all this research has often shown the continuing significance of national borders within Britain and the different perspectives on national identity that are sometimes evident across these borders, as much (if not more) than highlighting what is shared within a 'social union'. However, doubtless much social science has and continues to examine social and cultural phenomenon through an undifferentiated 'British' lens. This perspective sometimes also extends to a cultural studies approach that addresses some of the other purported aspects of a (British) social union (see e.g. chapters on film, television, music and sport in Morley and Robins (eds.) 2001).

4.4 While the analysis in this paper has been based largely on quantitative evidence concerning the articulation of national identities and political attitudes in Scotland, such analysis has itself consistently suggested that identifying as Scottish and/or British cannot be simply a political matter, and this confirms that being British may be as much if not more concerned with shared cultural, social and economic reference points rather than legal citizenship and a shared political structure. While the concept of social union has similar connotations, suggesting various ways in which British identification and sentiment may exist and be sustained somewhat independently of more formal political alliances and arrangements, perhaps its contemporary significance lies in its more explicit articulation as part of debates about political-constitutional futures in a radically changed post-referendum political landscape in Scotland and Britain^[11]. A further more novel and contemporary implication suggested by social union is that British and Scottish identities may continue to be complementary rather than competing, regardless of Scotland's political future (see also McCrone and Bechhofer 2015: 186-7). Indeed it has been suggested that the SNP, while being the primary advocates of the dissolution of the political (British) union, might even be described as the most faithful representatives of the continuing social union¹².

4.5 These observations can go some way to help us better understand the 'non-alignment' that continues to be evident between identities and attitudes to key political-constitutional questions in Scotland. Perhaps those who prioritise their Scottish over British identity do not also manifest the most obvious expressions of political Scottishness because of the extent to which they associate with the social union. Perhaps even those who state, in response to survey questions, that they do not feel nor would describe themselves as British still concede that they are British in some respects, whether they find this desirable or not. At the same time, a key finding in this paper is that it demonstrates recent evidence of closer 'alignment', suggesting the potential for more politicised national identities in Scotland, which may or may not indicate a weakening in people's attachment to social union.

4.6 A degree of caution is required in considering the extent to which these various factors might explain the results of the 2014 referendum and 2015 UK general election in Scotland. It seems likely that the willingness of a greater number of people who identified as exclusively or primarily Scottish to politicise these identities via support for an independent Scotland to some extent explains the higher than expected (prior to the campaign) Yes vote in the referendum. But at the same time at least some of these Yes voters might have envisaged a continuing identification with Britain regardless of the referendum outcome and there will have been some who identified as exclusively Scottish but who voted No for any number of reasons not related to national identity. Similarly, while closer alignment of identities and attitudes might go some way to explaining the SNP's overwhelming success in the 2015 UK election, it is likely that this success owes much to how the party was evaluated (in relation to its policies, organisation and leadership) in comparison to its main political rival in Scotland, the Labour Party. In other words, for many a vote for the SNP would have been motivated by political-ideological factors rather than being an expression of national identity.

4.7 Overall, national identities in Scotland, at least as they relate to political-constitutional matters, may be understood as a series of only partially overlapping and shifting constituencies, based on subjective national belonging, residence, political enfranchisement, political-constitutional attitudes, and people's understanding of and sense of affinity with a (British) social union. Although the outcome and aftermath of both the 2014 referendum in Scotland and the recent 2015 UK general election cannot be explained solely through reference to national identities, understanding these constituencies and how they interrelate is likely to be more important than ever for social scientists in this new and changing political context. Developing this understanding might also include further sociological investigation into the concept of social union, which has until now largely been only superficially articulated by politicians and has not been addressed explicitly and directly by social scientists. This would certainly involve research that goes beyond the employment of survey questions and includes a more qualitative, interactive dimension, building on the other related qualitative research outlined above. This would allow for the extent to which the concept of social union has resonance among the general public to be addressed more deeply and directly. To what extent do cross-border family and friendship networks, employment experiences (or aspirations) or particular cultural or sporting reference points contribute to a sense of Britishness and a sense of belonging to a shared social union? If these are important, then to what degree may they be separated from the key political attitudes that have been addressed in this paper, especially if these attitudes are indeed becoming more closely 'aligned' with people's subjective national identities? Perhaps most pertinently, would a 'social' Britishness be likely to endure in any future independent Scotland (and for how long), or would the end of the political union effectively and eventually mean the end of Britain?

Notes

¹Indeed, while taking a primarily legal perspective, Tierney himself recognises that one of the reasons that such issues are so legally complex is that they relate to identity markers and concepts that are sociological as well as legal in nature.

²'Qualifying citizens' are those who have leave to remain in the UK or do not require such leave. In contrast to Scotland's referendum, citizens of other EU states (excepting Ireland) are not permitted to vote in UK general elections (see http://www.electoralcommission.org.uk/faq/voting-and-registration/who-is-eligible-to-vote-at-a-uk-general-election).

³Based on 2011 Census data nearly 10% of Scotland's population were born in other parts of the UK and 3% in other EU countries (this does not include Croatia, which joined the EU in 2013). At the time of writing, detailed data on other Commonwealth citizens resident in Scotland were not available, but 2011 Census data indicates that around 1.2% of Scotland's population were born in the Commonwealth countries of Australia, Canada, India, New Zealand and Pakistan alone.

⁴The exact estimate is 5,847,000. The same source also estimates 3,570,000 people in the USA with 'Scotch-Irish' (as distinct from Irish) ancestry.

⁵This is an estimate based on a 20% sample of data.

⁶This term is commonly used by academic and other researchers. It reflects the fact that the question was first applied in a Scottish (and indeed UK) context by Luis Moreno (see e.g. Moreno 1988), although in fact a similar question had previously been used by other researchers in Spain (see e.g. Gunther et al. 1986).

⁷In both 2012 and 2013 the More Scottish than British category was marginally the modal category before rounding.

⁸It should be noted that there are only 70 cases in this group and hence the margin for error is much higher than for the other parties, particularly compared to SNP and Labour identifiers. But even with a much larger confidence interval the difference in the proportion with an equal dual identity compared to previous years is still significant.

⁹Indeed although not shown in the table, the data for 2013 broadly occupy a mid-point between the 2012 and 2014 figures for both parties.

¹⁰As well as social union, the former SNP leader and First Minister of Scotland, Alex Salmond, argued in the pre-referendum period that four other 'unions' would also endure following Scotland's political independence: the union of the crowns (i.e. the monarchy), the currency union, the defence union (via NATO membership) and Scotland's place in the European Union (Scottish Government 2013). These claims - especially regarding currency union and European Union - generated much controversy and debate in the referendum campaign.

¹¹Indeed, some have expressed scepticism about the extent to which the social and political (unions) can be neatly divorced (seeTickell 2012). Moreover, some conceptions of social union also imply that it could accommodate some more formal institutional and political dimensions such as diplomatic cooperation and shared social policies (Mooney 2013; Tickell ibid.)

¹²See Professor James Mitchell's Donaldson Lecture to the SNP party conference in October 2011: https://www.youtube.com/watch?v=oTUHHVlWQWY.

References