UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Separatist parties in central parliaments McGovern, Clare 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_september_clare_mcgovern.pdf [ 1.55MB ]
Metadata
JSON: 24-1.0166584.json
JSON-LD: 24-1.0166584-ld.json
RDF/XML (Pretty): 24-1.0166584-rdf.xml
RDF/JSON: 24-1.0166584-rdf.json
Turtle: 24-1.0166584-turtle.txt
N-Triples: 24-1.0166584-rdf-ntriples.txt
Original Record: 24-1.0166584-source.json
Full Text
24-1.0166584-fulltext.txt
Citation
24-1.0166584.ris

Full Text

SEPARATIST PARTIES IN CENTRAL PARLIAMENTS by  Clare McGovern  M.A., The University of British Columbia, 2004  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Political Science)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2015  © Clare McGovern, 2015ii Abstract This dissertation examines separatist political parties who compete in central elections and sit in central legislatures, focusing on the Scottish National Party, the 19th-century Irish Parliamentary Party and the Bloc Québécois. These cases represent an intriguing paradox: politicians who engage with the very institutions they wish to leave, using the existing political system to make a case for exit. Contrary to previous scholarly work on anti-system parties, my findings suggest that separatists do not always attack the political system. Indeed, they sometimes participate actively in institutions, despite being fundamentally opposed to the current constitutional set-up. These parties vary their electoral and legislative behaviour over time, in response to the potential for self-government and the structure of electoral competition. When self-government is taken off the immediate political agenda (for example after a failed referendum) I find that parties get more involved in parliamentary activities at the statewide level, beginning to use the legislature as a platform for their region's grievances. When self-government is imminent, separatists only participate in parliament if they are pivotal and therefore able to bargain over the terms of self-government. In election campaigns, separatists do not always focus on their region's grievances and their desire for independence. Grieving occurred most frequently when the separatist party was competing with the government party, as it was in its electoral interest to paint a negative picture of the region's fortunes, blaming this on the electoral rival. However, when competing with the official opposition party, separatists campaigned as the region's champion in the central legislature, claiming credit for services that their region has received.  In doing so, they tacitly supported the constitutional status quo, arguing that their region can prosper in the existing system. iii Preface  This dissertation is original, unpublished, independent work by the author, Clare McGovern.      iv Table of Contents  Abstract .................................................................................................................................................... ii Preface ..................................................................................................................................................... iii Table of Contents ................................................................................................................................... iv List of Tables ........................................................................................................................................ viii List of Figures ......................................................................................................................................... xi Acknowledgements................................................................................................................................ xii Dedication ............................................................................................................................................. xiv Chapter 1: Introduction ......................................................................................................................... 1 1.1 Why study separatist parties? .................................................................................................... 3 1.2 Separatist parties in the party politics literature ........................................................................ 6 1.3 An overview of the argument: the dilemmas facing separatist legislators ............................. 12 1.3.1 A framework for understanding the behaviour of individual legislators. ........................... 12 1.3.2 Applying the framework to separatist parties ..................................................................... 16 1.4 Testing the argument: cases and data ...................................................................................... 18 1.4.1 Country cases ...................................................................................................................... 19 1.4.2 Party cases ........................................................................................................................... 20 1.4.3 Data sources ........................................................................................................................ 22 1.5 Organization of the dissertation .............................................................................................. 24 Chapter 2: Theoretical framework ..................................................................................................... 26 2.1 Separatists in a central party system ....................................................................................... 26 2.1.1 Procedural focus .................................................................................................................. 27 2.1.2 Ideologically diverse vote base ........................................................................................... 28 v 2.1.3 Limited resources ................................................................................................................ 29 2.2 The dilemmas of a separatist party ......................................................................................... 31 2.2.1 The choices open to separatist parties: electoral strategy ................................................... 32 2.2.2 The choices open to separatist parties: institutional participation ...................................... 35 2.3 Resolving separatist dilemmas: constitutional, electoral and legislative opportunities .......... 38 2.3.1 The potential for separation ................................................................................................ 38 2.3.2 Electoral competition with the governing party .................................................................. 39 2.3.3 Electoral competition with the opposition party ................................................................. 40 2.3.4 Pivotal status ....................................................................................................................... 41 2.4 Interaction of the independent variables ................................................................................. 44 2.4.1 The impact on campaign strategy. ...................................................................................... 46 2.4.2 The impact on legislative participation ............................................................................... 49 2.4.3 Threats to causal inference .................................................................................................. 52 2.5 Predictions: how separatists behave over time ....................................................................... 54 Chapter 3: Electoral rhetoric and the prospects for self-government ............................................. 56 3.1 Case selection and data analysis ............................................................................................. 58 3.1.1 Defining “reasonable prospect” of self-government: selecting election campaigns ........... 58 3.1.2 Measuring election rhetoric: the dependent variable .......................................................... 61 3.2 Results for scenario 1: a pivotal separatist party competing against the opposition ............... 63 3.2.1 Do candidates explain their strategic position to voters? .................................................... 66 3.2.2 Campaign messages: pro- or anti-system? .......................................................................... 70 3.2.3 How does the IPP use its pivotal status? ............................................................................. 72 3.3 Results from scenario 2: a pivotal separatist party competing against the government ......... 79 3.4 Results from scenario 3: a non-pivotal separatist party competing against the opposition .... 85 vi 3.5 Results for scenario 4: a non-pivotal separatist party competing against the government ..... 92 3.5.1 The SNP in the Blair and Brown era ................................................................................... 92 3.5.2 The BQ in the Mulroney and Chrétien era. ......................................................................... 99 3.6 Conclusion ............................................................................................................................ 105 Chapter 4: How do separatist parties appeal to voters when self-government is off the agenda? ............................................................................................................................................................... 108 4.1 Case selection and data analysis ........................................................................................... 109 4.1.1 Selection of election campaigns ........................................................................................ 109 4.1.2 Measurement of electoral behaviour ................................................................................. 114 4.2 Testing the impact of electoral competition .......................................................................... 117 4.2.1 When the separatist party is not pivotal ............................................................................ 117 4.2.1.1 The SNP: 1983 v 2001 .............................................................................................. 117 4.2.1.2 The IPP: 1906 v 1909-10 .......................................................................................... 123 4.2.2 When the separatist party is pivotal .................................................................................. 127 4.3 Testing the impact of legislative status ................................................................................. 133 4.3.1 When separatists are competing with an opposition party ................................................ 134 4.3.1.1 The IPP 1909-10 v December 1910 .......................................................................... 134 4.3.2 When separatists are competing with a government party ............................................... 137 4.3.2.1 The SNP: 1979 v 2001 .............................................................................................. 137 4.3.2.2 The BQ: 2004 v 2005-06 and 2008 ........................................................................... 141 4.4 Conclusion ............................................................................................................................ 146 Chapter 5: When do separatists participate in the central parliament? ....................................... 149 5.1 Case selection ........................................................................................................................ 151 5.2 Participation in the central parliament. ................................................................................. 155 vii 5.3 The Bloc Québécois in the 35th Parliament: the impact of the sovereignty referendum ...... 158 5.4 The SNP and the Callaghan government: the impact of the devolution referendum ........... 166 5.5 The Bloc Québécois in the 37th and 38th Parliaments: the impact of minority government. 172 5.6 Conclusion ............................................................................................................................ 178 Chapter 6: Conclusion ........................................................................................................................ 181 6.1 Key findings .......................................................................................................................... 181 6.2 Contributions ......................................................................................................................... 185 6.3 Limitations of the research .................................................................................................... 190 6.4 Future research ...................................................................................................................... 193 References ............................................................................................................................................ 196 Appendices ........................................................................................................................................... 209 Appendix A Case selection ............................................................................................................... 209 Appendix B Coding electoral rhetoric. ............................................................................................. 216 B.1 Data sources ...................................................................................................................... 216 B.2 Categorization of electoral rhetoric ................................................................................... 218 B.3 Coding electoral rhetoric ................................................................................................... 222 Appendix C Additional break-downs of electoral rhetoric data: Chapter 3 cases. ........................... 226  viii List of Tables  Table 2.1 The electoral and legislative choices of a separatist party ...................................................... 37 Table 2.2 Separatist campaign behaviour: the interacting effects of (1) prospects of self-government, 46 Table 2.3 Legislative participation: the interacting effects of (1) prospects of self-government and .... 51 Table 3.1 Categorizing campaign messages, IPP 1886 versus 1895 ...................................................... 71 Table 3.2 Categorizing campaign messages: SNP, October 1974 versus 1979 ...................................... 85 Table 3.3 Categorizing campaign messages: SNP, February 1974 versus 1983 .................................... 88 Table 3.4 Categorizing campaign messages: SNP, 2005 versus 2010.................................................... 95 Table 3.5 Categorizing campaign messages: BQ, 1993 versus 1997 ................................................... 103 Table 4.1 Descriptive claims in SNP campaigns: who gets the blame and who gets the credit? ......... 118 Table 4.2 Descriptive claims and attributions made in the 1906 and 1909-10 IPP manifesto ............. 125 Table 4.3 Negative descriptive claims in SNP pamphlets, Aberdeen, 1979 ......................................... 128 Table 4.4 Positive descriptive claims in SNP pamphlets, Aberdeen 1979 ........................................... 130 Table 4.5 Descriptive claims made in IPP electoral campaign materials ............................................. 135 Table 4.6 Negative descriptive claims and attributions made in SNP electoral materials.................... 139 Table 4.7 Negative descriptive claims and attributions in BQ electoral campaign materials .............. 142 Table 4.8 Positive descriptive claims and attributions made in BQ electoral materials ....................... 145 Table 4.9 Predictions and conclusions: testing the impact of electoral competitor .............................. 147 Table 4.10 Predictions and conclusions: testing the impact of legislative status ................................. 148 Table 5.1 Total number of Private Members' Bills and Motions submitted in the Canadian Parliament ............................................................................................................................................................... 156 Table 5.2 Private Members' Bills in the 1993-97 Canadian Parliament ............................................... 160 Table 5.3 Private Members' Motions in the 1993-97 Canadian Parliament ......................................... 161 ix Table 5.4 BQ's Private Members' Bills by subject, 1993-97 Parliament .............................................. 164 Table 5.5 Private Members' Bills on Québécois or Francophone issues, 1993-97 Parliament ............ 165 Table 5.6 All parties' use of questions on House of Commons business, 1977 and 1979, UK Parliament ............................................................................................................................................................... 168 Table 5.7 The SNP's questions on House of Commons business, 1977 and 1979, UK Parliament ..... 170 Table 5.8 Private Members' Bills in the Canadian Parliament, 2004-05. ............................................. 173 Table 5.9 Private Members' Motions in the Canadian Parliament, 2004-05 ........................................ 174 Table 5.10 BQ Private Members' Bills by subject, 2000-04 and 2004-06 Parliaments ........................ 175 Table 5.11 BQ Private Members' Motions by subject, 2000-04 and 2004-06 Parliaments .................. 175 Table 5.12 Private Members' Bills on institutional reform, 2000-04 and 2005-06 Parliaments........... 176 Table 5.13 Private Members' Motions on institutional reform, 2000-04 and 2005-06 Parliaments ..... 176  Table A. 1 Elections examined in Chapter 3 ........................................................................................ 209 Table A. 2 Elections examined in Chapter 4 ........................................................................................ 212 Table A. 3 Parliamentary sessions examined in Chapter 5 ................................................................... 215  Table C. 1 Categorizing campaign messages: IPP, 1886 ...................................................................... 226 Table C. 2 Categorizing campaign messages: IPP, 1895. ..................................................................... 226 Table C. 3 Categorizing campaign messages: SNP, February 1974. .................................................... 227 Table C. 4 Categorizing campaign messages: SNP, October 1974 ...................................................... 227 Table C. 5 Categorizing campaign messages: SNP, 1979 .................................................................... 228 Table C. 6 Categorizing campaign messages: SNP, 1983. ................................................................... 228 Table C. 7 Categorizing campaign messages: SNP 2005 ..................................................................... 229 Table C. 8 Categorizing campaign messages: SNP 2010 ..................................................................... 229 Table C. 9 Categorizing campaign messages: BQ 1991 and 1993 ....................................................... 230 x Table C. 10 Categorizing campaign messages: BQ 1997 ..................................................................... 230 xi List of Figures  Figure 2.1 Separatist behavior: constitutional, electoral and legislative opportunities .......................... 45 Figure 3.1 Selecting election campaigns ................................................................................................. 60 Figure 3.2 Categorizing campaign messages .......................................................................................... 62 Figure 4.1 Paired elections, varying electoral competitor and keeping legislative status constant ..... 110 Figure 4.2 Paired elections, varying legislative status and holding electoral competitor constant ...... 112 Figure 4.3 Categorizing campaign messages ........................................................................................ 116 Figure 5.1 Paired cases, varying prospects of self-government; holding pivotal status constant ......... 154 Figure 5.2 Paired cases, varying pivotal status; holding prospects of self-government constant ......... 155 Figure B.1 Categorizing campaign messages ....................................................................................... 223              xii Acknowledgements Writing these acknowledgements is the final part of this dissertation – and also the most satisfying as there are so many people to thank. First, my adviser Chris Kam, without whom I would not have finished  - many thanks for your perseverance, insight and for introducing me to C19th electoral addresses (reading those was definitely my favourite part of the research). Alan Jacobs not only got me excited about methods, I have lost count of the all his thoughtful comments over the years that got to the heart of what I was trying to say. Paul Quirk has been a fantastic source of advice and encouragement on both the dissertation itself and where it goes next. I’ve also had incredibly helpful advice from other UBC faculty, on my research and on the overall doctoral process: my thanks to Gerry Baier, Colin Campbell, Ken Carty, Fred Cutler, Diane Mauzy, Ben Nyblade, Campbell Sharman and Lisa Sundstrom. Finally, nothing would get done in our department without Dory Urbano and Josephine Calazan – you have both helped me out many times and been great friends.   I am grateful for the financial support provided at various points in the programme by the UBC University Graduate Fellowship and the C.D. Howe Graduate Fellowship. I also owe thanks to the staff at the University of Aberdeen Library for sending copies of their collection of Scottish National Party leaflets. Particular thanks to Elise Leclerc-Gagné, for your translation expertise, your explanations of uniquely Québécois terms and, most of all, for putting up with a conquering Brit.   There is a long list of friends who gave advice and feedback on the research, the draft or the defence presentation, and, most importantly, told me many times to keep going. Thanks to: Afsoun Afsahi for a shared HP and BtVS obsession; Nathan Allen for being my favourite boss and critic; Solanna Anderson for being my fellow Academics Anonymous member; Sarah Banting and Tasha Riley for being my family in Vancouver; Kaija Belfy Munroe for much-needed perspective and for luring me to Squamish; xiii Katie Boothe for methods and movies chats and because I’ve known you long enough to remember our “Who is this Senator Obama?” discussions; Adam Bower for being my fellow LFC fan in the wilderness; Katrina Chapelas for keeping us all sane and for general encouragement of shenanigans; Ella Connor for many things over many years and encouraging me to keep going at a crucial point; Nick Dragojlovic for over a decade (!) of unfailingly good advice; Sean Gray for the football and age-related mockery; Natasha Krowchuk for travelling fun and BC wine; John McAndrews for our joint love of comparing parliaments; Sophia Ongley for inter-disciplinary conversations on grad life; Avery Poole for late-night pho and teaching debriefs; Beth Schwartz for applying your steel-trap mind to my drafts and for a mutual appreciation of sloths; Aim Sinpeng for always cheering me up; John Voortman for being my partner in late night caffeinated procrastination; Azur Tyer for a long friendship and brilliant coffee-dates; James Tyer for being my fellow ex-pat; Liz Victor for transatlantic PhD comparisons and for sending me off on a 7am snail hunt with your son; Daniel Westlake for many fascinating conversations in the grad room and Carla Winston for Dr. Who appreciation.   Most of all, I owe a huge amount to my family. To my parents Tegwen and Ged, for awaking the fascination with politics in the first place and for always being enthusiastic about my meandering career path. To my brothers, Chris, Martin and David, for a near-lifetime of companionship, and more importantly, mockery. To my sisters-in-law, Jo and Steph, I’ve always appreciated your friendship; thanks for evening out the family gender balance! To my wonderful nephews and niece: Matthew for the incessant Harry Potter riddles and superhero stories, Olivia for introducing me to every dog on Merseyside and my very new nephew Josh for being born. Finally, this dissertation is dedicated to my grandmothers, May McGovern and Joyce Williams – two extraordinary women who showed deep courage as young women and a lifelong commitment to family and to women’s education. My love and thanks to you both. xiv Dedication  In memory of my grandmothers, May McGovern and Joyce Williams.  1 Chapter 1: Introduction When first elected to the Westminster Parliament in 1918, Sinn Féin, the Irish Nationalist party, refused to take its seats, instead establishing an alternative institution in Dublin.1  After the partition of Ireland, Sinn Féin continued to run in Westminster elections in Northern Ireland. However it never deviated from its original boycott of the British Parliament, as taking seats would entail an oath of allegiance to the British Crown.2  At first sight, Sinn Féin appears to be the quintessential example of a separatist party: running for election in order to protest against the state, but refusing to actually participate in the state’s institutions.   However, such disengagement is unusual for separatist political parties.3 Young and Bélanger survey all separatist parties that have won seats in established democracies since 1970 (490, 2008; based on Sorens, 2004, 2005). Of these, Sinn Féin is the only party to have never participated in legislative activities. At the other end of the spectrum, two parties, the Lega Nord and Convergència i Unió, (CiU) joined governments at the statewide level, with their leaders accepting ministerial office.4    In between these two extremes are many subtler distinctions. The Scottish National Party (SNP) seeks full independence for Scotland and has held seats at the Westminster Parliament since 1967.5  Its members swear oaths to the British Queen and participate in some debates and votes, but have relatively low turnout levels; in the 2007/08 parliamentary year, for example, they took part in only 35% of legislative votes (McGovern, 2009, 18). Compare this to the Bloc Québécois (BQ)’s far higher                                                  1 This was to become the Irish legislature after independence. 2 See Hansard (UK), 20 April 2006, Cols 275-276 (exchange between Stephen Pound MP and Lady Sylvia Hermon, MP) for a discussion of Sinn Féin’s actions and motivations. 3 Defined as those whose leadership has made a definitive and recent statement that the party seeks self-government for its region.  4 These parties sought autonomy for Northern Italy and Catalonia respectively. 5 This question of independence was put to a referendum of Scottish voters in September 2014; 45% of voters supported independence, 55% opposed it. (BBC Online, 2014) 2 levels of participation in the Canadian Parliament (54% of legislative votes in the same parliamentary year; ibid). Furthermore, the Bloc took on the role of the Official Opposition in the 1993-97 Parliament, despite the fact that this period coincided with a referendum in which it was campaigning for Québec’s political and economic sovereignty.  More recently, the BQ pledged to actively help a potential government maintain the confidence of the House of Commons (Valpy, 2009).   The Irish Parliamentary Party (IPP), which sought Irish Home Rule during the late 19th and early 20th centuries, took this one step further. In sharp contrast to its electoral rival Sinn Féin, it not only sat in the UK House of Commons, but actually supported several British governments. While it did not go as far as the Lega Nord and sit in the Cabinet, it worked closely with government ministers. Co-operation was not simply on matters of confidence, nor on policy concerning Ireland alone, but on constitutional reforms to the entire United Kingdom. In 1910-11, the IPP helped the Asquith government pass the Parliament Act which reduced the veto power of the House of Lords (Wicks, 2006; Lyons, 1973).   Among separatist parties, then, there is a spectrum of engagement, ranging from a complete rejection of the central political system (Sinn Féin) to inclusion in the government (Lega Nord and the CiU). These cases point to an intriguing paradox: politicians who expend time and energy engaging with the very institution they want to leave, using the institution’s own rules and procedures to make a case for exit. Furthermore, these parties have continued to run for office, win seats and participate in the central legislature over several decades: being a part of the political system, yet apparently convincing their voters that the status quo is unacceptable. This dissertation explains when and how separatist parties engage in the political system, showing that they are neither inherently co-operative nor intransigent. Rather, their behaviour varies over time in response to electoral and institutional incentives and to the likelihood of separation itself.  3 This chapter sets out the case for studying separatist parties, showing how existing literature on political parties fails to address important questions about the behaviour of separatist legislators. I then outline a theory of separatist politicians’ preferences and the dilemmas they face, from which I draw predictions of separatist behaviour under varying electoral and legislative conditions. Finally, I explain how the dissertation tests these predictions using the cases of the Scottish National Party, the Irish Parliamentary Party and the Bloc Québécois.  1.1 Why study separatist parties? Research on the behaviour of separatist parties develops our understanding of a widespread empirical phenomenon and complements existing literature on political parties. However, it also raises questions of wider interest on the peaceful management of fundamental tensions within a political system and on the creation and maintenance of political institutions.   Parties which focus on the identity and demands of a sub-national region are so common an occurrence (Sorens, 2012; Swenden and Maddens, 2009; Young and Bélanger, 2008) that Gunther and Diamond include “pluralist nationalist” parties in their typology of political parties (2003, 172). Gunther and Diamond argue that if these parties are to be electorally successful they need to foster a sense of identity – one that is distinct from the rest of the state. The authors conclude that such parties “almost by definition . . . involve a demand for some level of territorial self-government” (2003, 180-181).   Not only are separatist parties common, their very existence tells us something about inter-regional relations within a country. These parties’ electoral success suggests a substantial incidence of support for secession among the region's voters, or at the very least a desire to bring governmental decisions closer to home. How such parties interact with the political system is normatively important for two 4 reasons. First, if there is no outlet for the democratic consideration of demands for self-government, there is a danger the secessionist movement will turn to extra-constitutional avenues. This risks violence and the destabilization of the entire country (Crick, 1971, 40). Second (and conversely), the literature on “anti-system parties” suggests that it is the parties themselves (rather than the state’s response) that may be harmful.6   Keren argues that by definition such parties try to undermine the legitimacy of the system. Their rhetoric only points to the failures of political institutions rather than their successes. Although this is expected behaviour for anti-system parties, he argues it is also dangerous because it reduces public confidence - not just in the government of the day, but in the entire system of representation (2000).   These purported dangers point to a need for empirical study of the behaviour of anti-system parties. Do they always attack the overall system and all other parties in order to achieve their goals? Are they able to function in the political system as electoral candidates, representatives and legislators and to pursue their agenda of self-government?  Separatist parties are a particularly interesting type of anti-system party as they are not seeking to transform or destroy the existing constitutional structure, but simply to exit from it. Studying how such parties deal with central institutions in the meantime gives us an insight into the effect of time horizons on behaviour. The work of both historical and rational choice institutionalists stresses that time matters in institutions (see for example Jacobs, 2011; Pierson, 2000; 2004; Shepsle 1989). Mainstream parties take positions in elections and in policy debates while expecting to be a part of the political system for the foreseeable future. Even if they fail in the short-term, they still have a chance of achieving what they really want at another point in the future. Furthermore the choices political actors make when                                                  6 Sartori defines these as parties with goals that are incompatible with the state’s existing constitutional and territorial structure (1971, 33). 5 competing or co-operating with each other reflect the expectation that they will continue to interact with the same players in the future: reputations therefore matter (see e.g. Axelrod and Hamilton 1981; Axelrod 1986, 1997). Finally, political parties may expend considerable time and energy setting up and maintaining a forum for interest aggregation and decision-making (i.e. a legislature) if they expect to be using that institution for the foreseeable future.   How is this different for a party whose raison d’être is to leave the political system? Such a party has to make inter-temporal trade-offs, between winning votes at the next election and preparing for a separate political system in the future. This suggests that the party’s members will not care about the rules and norms governing party interactions in Parliament as they do not want it to carry on making decisions about their region. So, how will separatists engage with mainstream politicians if they do not intend to be interacting with them indefinitely into the future, and does this vary with the likelihood of separation?   A common response to inter-regional rivalries has been federal constitutions or the devolution of power to regional bodies, to prevent tensions developing into full-blown demands for independence.7  Indeed, several of the parties already mentioned also participate in sub-national legislatures and have led governments at this level. However, this dissertation focuses on the behaviour of separatists within the central legislature, not the sub-national institutions in its own region. Secession is a process which affects the whole country: while a regional body may unilaterally declare its independence, in reality this is unlikely to be effective (or peaceful) without negotiations with the central body.  Debates about separatism or autonomy will still therefore take place at the central level, even in federal states or states where substantial devolution of power has already occurred. Furthermore, to the separatist party, a                                                  7 See Burgess, 2006 for an overview. 6 regional legislature is a prototype for a potential parliament in an independent state. In other words, its members do intend to continue to use the regional institution (or one like it) and to interact with other regional parties in the future. Their behaviour in the regional legislature may therefore be a very misleading indicator of how they act in the central legislature.   1.2 Separatist parties in the party politics literature The phenomenon of separatism has been studied by scholars of both intra-state conflict and political parties, but this research has not addressed the cross-national phenomenon of separatist parties which participate in central legislatures. This is because existing work has either focused on one party in particular or on activities and actors outside the legislature, or it has failed to test its predictions systematically against the legislative record. It therefore has not built generalizable propositions as to how separatist parties interact with the institutions they want to leave, nor tested these across a range of cases.   Given the parliamentary system’s fusion of the executive and the legislature, it is unsurprising that comparative research on parliamentary behaviour has tended to revolve around the government. Such work has covered the formation of government, its structure and survival, scrutiny of that government’s performance and its success in completing its legislative agenda (see, e.g. Kam and Indriđason, 2005; Laver and Shepsle, 1996; Martin and Stevenson, 2001; Tsebelis, 2002; Warwick, 1994, 2006).  This in turn has led to a focus on contenders for government: parties likely to win a majority of the seats in the legislature or to be pivotal.8  Given the limited territory in which they can compete, separatists will only rarely have enough legislators to fall into this category – and even when they do, Warwick’s                                                  8 Researchers’ prioritization of such parties is borne out by the Comparative Manifesto Project’s decision to only gather manifesto data for large or pivotal parties. (Budge et al, 2001). 7 research suggests a widespread reluctance to include such parties in coalition governments (1994, 4; 46-7).   An exception is the literature focusing on voting patterns in the legislature, where researchers include voting data for smaller parties in their analysis. However, the repeated finding that government versus opposition is the dominant voting dimension in parliamentary systems (Dewan and Spirling, 2011; Godbout and Høyland, 2011; Hix and Noury, 2014) means that the actions of smaller parties tend to be subsumed into those of larger parties: they are either in the government and vote accordingly or they follow the lead of the main opposition parties. There is therefore little incentive to theorize the preferences of small party leaders and the legislative actions of small party members as distinct from those of their more influential peers.   Accordingly, where scholars have studied small parties, the predominant research question has not been about their legislative behaviour, but whether they get into the legislature in the first place. Lines of research have included the ways in which electoral systems and voter behaviour can disadvantage smaller parties (Cox, 1997; Ezrow 2010) and small parties’ use of niche issues to overcome such barriers to entry (Bélanger and Meguid, 2008; Meguid, 2005; 2008). In fact, Meguid’s work includes a case study of the SNP. However, her focus on electoral competition means that the central questions are whether a small party can use a new issue to take votes from mainstream parties and how the larger parties respond. In her account, therefore, the issue of Scottish independence is treated as being analogous to environmental protection and immigration, which have been the springboards for other niche parties. However, viewed through a legislative lens, separatist parties are very different. Green parties and far-right, anti-immigration parties (Meguid’s other cases) are challenging existing policy priorities, but not the entire constitutional structure. They do not face the paradox of the separatist party 8 in a state-wide legislature: the reward for their electoral success is participation in the very institution they seek to leave.  Separatist parties therefore face a dilemma between engaging with the political system in order to publicize and negotiate their constitutional demands and boycotting political institutions to underline their rejection of the current system's legitimacy.   So, if the comparative literature on legislatures and parties has not engaged with the unique dilemma faced by separatist parties, what about the literature on the phenomenon of separatism itself? Much of this research seeks to understand why there is a demand for secession in the first place. It therefore asks why certain ethnic identities, ancestries or histories are particularly salient and why they lead to a desire for self-government (see, e.g. Ayres and Saideman, 2000; Fearon and Laitin, 2000; Horowitz, 1981, 1985; Laitin 2000; Sorens 2004, 2005; Wood, 1981). Research answering these questions is useful for explaining why separatist parties have electoral appeal but not for explaining the behaviour of their elected representatives.    The conflict management strand of the separatism literature focuses on the state’s response to separatism. Existing work either explains the bargaining process between the government and separatists (see, e.g. Cunningham 2006; Gallagher Cunningham 2011; Sorens 2012; Walter 2006a) or surveys constitutional structures which can accommodate sub-national demands while keeping the country together (see, e.g. Horowitz, 1993, 2002, 2003; Kymlicka, 1998; Lijphart, 1977; Nordlinger, 1972; Rudolph and Thompson, 1985). As Young and Bélanger note, much of this research is based on cases where separatism is intertwined with civil conflict (2008, 491). In this work, the actors are not necessarily political parties and the arena is not always a legislature or an electoral campaign. The separatists studied are broad social movements which put pressure on the government through a range of extra-constitutional, sometimes violent, means.  9 Conversely, the actor seeking to prevent this separation is conceptualized as the unitary state, not multiple political parties which also compete among themselves (Gallagher Cunningham, 2011, 275). Any focus on the established political system is at a similarly high level of abstraction, focusing on the division of powers between various institutions or the routes of entry to these institutions, and the knock-on effect on the power of different ethno-national groups.  This strand of the conflict management literature therefore does not deal with the day to day activities of separatists who operate within relatively stable political systems.   However, there have been a number of scholars who have considered the impact of “anti-system parties”. These are defined as parties which oppose not merely the government of the day but the constitutional structure itself and for whom changing or leaving the political system is a central part of their ideological identity (Sartori, 1971, 33; Schedler 1996, 307). This definition clearly includes separatists. As noted in the previous section, these authors make clear predictions about the legislative behaviour of such parties: that they will either disengage from the institution altogether or will obstruct routine business, employing rhetoric that is unremittingly hostile to the existing system and to mainstream parties (Schedler, 1996; Keren, 2000).   These claims leave several questions open. First, there is a lack of systematic, cross-national empirical work to back up these claims. There have been studies of individual anti-system parties’ legislative activities, focusing on the topics they raise in Parliament (Gold 2003; Tambini, 2001; Young and Bélanger, 2008). There have also been cross-system studies of the quality of deliberation found in the legislative exchanges of mainstream parties (Steiner et al, 2004). However, there has been no work which compares anti-system to pro-system parties, or indeed one anti-system party to another.    10 Existing literature on parties and legislatures therefore cannot tell us if anti-system parties as a category participate in or obstruct parliamentary activities, nor how their rhetoric compares to that of mainstream parties. This lack of comparative empirical work means that existing commentary on anti-system parties assumes that particular types of behaviour follows logically from anti-system objectives. However, this claim ignores the possibility of two-way interaction: it argues that anti-system parties will (negatively) impact the institutions they are in but does not consider the possibility that institutions may affect politicians who participate in them.9  Anti-system parties are thus conceptualized as static actors, pre-determined to behave in a particular way and unaffected by changing environmental factors or interaction with other parties – and these assumptions have not been empirically tested.   Furthermore, just as Meguid’s work treats the SNP as analogous to other niche-issue parties, the anti-system literature does not distinguish between types of anti-system parties. For countrywide parties that reject the current political system, the remedy is to destroy or transform the constitutional order. In order to do so, they seek to persuade voters across the country that the current system is corrupt, oppressive and beyond repair. By contrast, a regionally-based group which has concluded that the status quo is unacceptable has another option: it can leave. Separatists do not have to overthrow the entire constitutional structure. They therefore have no interest in reducing general confidence in the political system among voters outside their own region.                                                    9 Both rational choice and historical institutionalist scholars argue that institutions can affect the behaviour of individuals who operate within them, whether through structured incentives (see, e.g. North, 1990; Shepsle, 1989) or through norms and altered expectations (see Steinmo, Thelen and Longstreth, 1992 for an overview). Docherty argues that this explains the actions of newly elected Members of the 1993-97 Canadian Parliament, many of whom belonged to parties which were critical of the existing political system, but who changed their position over the course of the Parliament (1997).  11 Research on particular separatist parties supports the argument that they do not automatically seek to undermine the legislature. Rather, behaviour varies across parties.10 However, as these case studies focus on specific separatist parties in isolation, they do not cast much light on why separatist parties differ from one another. Insight into their legislative or electoral behaviour is gleaned from a few high profile examples concerning either constitutional reform or government formation, rather than a systematic examination of day to day activities over the full range of issues (see e.g. Giordano, 1999; Guibernau, 2000; Lloyd, 1998, Lynch, 2002; Lyons, 1973; MacCormick, 2000; McAllister, 2001). An exception is Young and Bélanger’s study of the BQ’s use of oral questions in the Canadian House of Commons (2008). While providing valuable insight into how the BQ’s attention to particular issues varies over time, this study again focuses on the party in isolation. The questions asked by the BQ MPs were under the control of the questioner (subject to parliamentary rules of procedure) and therefore do not capture strategic interaction with other parties in the way that debates and electoral campaigns do. Finally, where literature on separatist parties does focus on  relations with mainstream parties, this is analyzed through the lens of electoral competition alone, thus ignoring what separatist politicians do once in the legislature (Anderson and Gidengil, 2002; Fusaro, 1979; Galatas, 2004; Levy 1995; McAllister, 2001; Meguid 2005; 2008).   As this overview of the relevant literature shows, we are left without an account of separatist parties as a distinct, cross-national phenomenon: actors which are elected to an institution on the basis that they want to leave it. There are developed theories of parliamentary party behaviour that do not address the particular constraints facing separatist parties. Conversely, there is much empirical detail on how specific separatist parties have behaved at certain moments in time, but no generalizable theory on their legislative and electoral behaviour. As a result, we are left with the questions raised in the previous                                                  10 See, for example, Young and Bélanger’s classification of separatist parties as either “democratic” or “obstructionist”, based on an overview of this literature (2008, 490). 12 section: how do politicians act when they have no long-term commitment to an institution, and do they achieve any of their goals by participating in that institution in the meantime?  1.3 An overview of the argument: the dilemmas facing separatist legislators  I argue that, like any other elected representative, a separatist politician has both primary and secondary objectives. At any point in time, however, she has to make trade-offs between achieving the goals she values most highly and goals which are of lower value but easier to achieve. Where the separatist differs from other politicians is the systemic focus of these objectives: secondary goals are pursued within the current political system in order to fight for her primary goal, which is the establishment of a new, separate political system. The central dilemma is therefore not simply a trade-off between different time horizons, but between two political systems. Does she focus on the current system (which she opposes but which is the location of political power) or the establishment of an ideal but as yet hypothetical electoral and legislative system?  This section begins by setting out a general framework for assessing an individual legislator's objectives and how these are aggregated to the collective actions of a political party. I then apply this argument to the specific case of a separatist party, explaining how its preferences and trade-offs differ from those of mainstream parties.   1.3.1 A framework for understanding the behaviour of individual legislators.  This research is based on the assumption that individual legislators act purposely in order to achieve multiple aims. A legislator’s primary aims are the ones on which she places the highest value. They reflect the reason she chose to go into politics and fall into two broad groups: achieving outcomes across a range of policy issue areas and receiving the perks of elected office, such as financial 13 compensation and status (Müller and Strøm, 1999). Individuals will of course vary in terms of whether policy trumps perks or vice versa, and which policy issue or which type of perk is most important. The value of the perks themselves will depend on the office held by that individual. Pay, benefits and status are likely to be highest for members of the government and lowest for backbench members of small opposition parties, with separatist legislators likely to belong to the latter category.   Legislators also have secondary aims, valued as a means to obtain policy outcomes or perks. Achieving elected office is an instrumental aim as, separate from the perks it provides, it offers the legislator a means by which to influence policy outcomes. A government minister has the authority to direct bureaucrats within her department, a budget, influence over Cabinet discussions and privileged access to legislative time, all of which can be used to pursue a range of policy goals. Even the newest backbencher in the smallest party gains resources which can be used to exert some degree of influence over policy: the right to vote on public expenditure and legislation, to propose amendments to legislation, to ask questions of government ministers and to make speeches in a public and authoritative forum. Furthermore, these legislative activities give politicians a platform from which they can gain media coverage for their views, and thus attempt to shift public opinion in their favour.   So, office is useful for its influence over policy and public opinion and can be valued in its own right for its status and perks. To achieve office in the first place, a legislator needs to win votes. She also needs institutional rules which allow her to access the tools to influence policy: the right to vote, offer amendments, speak, sit on committees and ask questions. Furthermore, these institutional rules can directly affect the value of the perks she receives from elected office: whether materially (for example, pension schemes and rules for claiming expenses) or symbolically (for example, rules of speaking order help to convey the status of the speaker).  14 This relationship between policy, office, votes and institutional rules is not static. At any one point in time, legislators have already won the votes to attain their current office, but must look forward to the next election if they are to retain their position, or improve upon it. Their legislative activities and the associated policy outcomes help them build a public profile on which to campaign. However, some activities may actually hinder a re-election attempt. A legislator’s primary policy preferences may conflict with public opinion or the material interests of her constituents. Speaking or voting in favour of such unpopular policies would therefore mean prioritizing her primary goals over her secondary goals to win votes, which are needed in order to have any impact on policy in the future. The preference which prevails will depend partly on whether the policy issue in question is particularly salient to the legislator.   Institutional rules are a means of achieving policy outcomes and building electoral appeal. Legislators therefore have an interest in involving themselves in discussions about reform of the institution, whether to change rules in order to increase their debate, committee and vote opportunities, or to defend rules which are already in their favour. Such institutional change may include either formal changes to rules or informal adjustments of conventions about appropriate legislative behaviour.11 A legislator attempts to affect the former by explicit participation in committee work, debates and votes on the subject and can influence conventions simply by day to day participation in general legislative activities.12   So, at any particular point, a legislator has three different considerations. Any particular legislative activity, such as casting a vote a certain way or making a speech, may have an impact on a policy                                                  11 See North, 1990 on formal and informal institutions.  12 See Docherty, 1997 for an empirical example of episodes where legislators have consciously challenged prevailing legislative norms.  15 outcome (i.e. a bill passes or not; an amendment is made or not). It may also affect her electoral chances. Even if the legislator loses the vote, or a speech does not affect the final text of legislation, the public attention garnered may win or lose her votes at the next election. Over the long term, her activities may shape the institution itself – either directly if she sits on a committee which is explicitly considering reform or indirectly by contributing to norms about how legislators behave.   This general framework has been set out in terms of individual legislators. However, recognizing that nearly all legislators are members of parties which have collective aims does not change the structure of the argument. Party membership provides politicians with brand recognition and financial and professional support for their election campaigns, while allowing them to have a much greater collective influence on legislative activities than they could achieve on their own.13 In return, legislators commit themselves to the party policy programme which aggregates the preferences of its supporters and they submit to party discipline over their activities.14  In other words, MPs agree to pursue their party leadership’s goals in return for a greater chance of being able to achieve those goals and on the understanding that those goals will not stray too far from their own, individual preferences.15 By virtue of their office as MPs, however, legislators still have the opportunity to pursue their own goals, provided that these do not conflict with those of the leadership: either by being substantively different, or by taking time away from activities crucial to the leaders’ aims (i.e. voting the party line and making speeches supporting the party and attacking opponents).                                                    13 See Cox, 1987, 1997 and 2005 for the electoral and legislative benefits of party membership.  14 See Kam, 2009 for a model of how parties provide incentives for MPs to follow the leadership’s preferences rather than their own.  15 The degree of party discipline can therefore be seen as a function of, first, the gap between the party leaders’ goals and those of backbenchers and, second, the leader’s success in achieving promised aims (such as winning an election, delivering a particular policy or achieving constitutional change).  16 1.3.2 Applying the framework to separatist parties For the leaders of a separatist party, the most highly valued preference is to set up a new institution so that their region can be self-governing, either as an autonomous unit within the existing state or as a completely new country. All other goals are secondary to this objective. In the meantime, though, leaders are operating in the central legislature in order to further the ultimate goal of autonomy or separation. If circumstances permit, they may support or even join a government in exchange for concessions on self-government.16 However, as noted in section 1.2, such opportunities are likely to be rare for separatists. How else, then, can they pursue their agenda by participating in the central legislature? First, separatists can use the institution as a platform to increase their party’s profile with the aim of increasing their vote at the next election. An increased share of the public vote allows them to claim that public opinion is moving in favour of separation, while winning more seats increases their chances of being pivotal and thus able to extract concessions on self-government.  Second, they can use the institution to influence other policy areas which affect their region, either to appeal to voters or to ease a transition to autonomy.   While they are still sitting in the central legislature and competing in central electoral campaigns, separatists have to pay attention to how these legislative and electoral institutions operate, to ensure they are able to ask questions, put down motions, sit on legislative committees and campaign effectively. Such tools are crucial for the aims of increasing the party’s public profile and influencing policy.  Separatist parties therefore engage in similar activities to other politicians: electoral campaigns, scrutiny of policy and government performance and institutional maintenance and change. However, each of these activities throws up particular dilemmas for separatists. First, do they prioritize short and                                                  16 We can assume that the perks of government office are of low utility to these politicians. If they valued being in government for its own sake, then they would have joined a mainstream party which offered a much better prospect of achieving this goal.   17 medium term secondary aims (i.e. votes and ensuring that the central institutions give them the opportunity to represent their region's interests) or the long term primary aim of self-government? Second, do they pursue those aims by disengaging from, actively attacking or co-operating with the existing political system?  In electoral terms, separatists can position themselves as the region’s champion in the central legislature: there to pressure the government for investment and public services and to raise the region’s historical, cultural or economic grievances. However, claiming credit for any benefits that the region receives means sharing that credit with the governing party (whose ministers have written the budget and whose members have voted for it). It also entails a recognition that the region does receive benefits under the current political system, thus undermining the case for separatism. Conversely, if separatists concentrate on being a party of grievance which constantly denounces institutions, all other parties and the rest of the population, this may alienate pro-system voters and make it hard to share in the credit for any government decision that does benefit the region.    On questions of policy, the party leadership’s first preference is that any legislation affecting its region be made in another, hypothetical institution. In the meantime, should party members refuse to participate in substantive policy discussions and thus forgo influence over legislation which will affect their region and voters? Or should they engage in debate in policy areas other than self-government and risk diluting their message and splitting their support base?   Finally, how much attention should leaders pay to the day-to-day operation of the central legislature? If there are rules and practices which disadvantage their members, should they devote time to trying to 18 change these, so that they have a platform, or does expending any time and energy on an institution they aim to leave merely make them look hypocritical?   This dissertation argues that these trade-offs vary in response to incentives in the party’s environment. Therefore, contrary to the claims of the anti-system party literature, separatist parties cannot simply be classified as inherently either co-operative or intransigent. These incentives revolve around opportunities for the separatists to pursue their preferences. First, party leaders are only likely to prioritize policy, votes or institutional change respectively if they expect to gain benefits from them and to continue to enjoy these benefits for a considerable degree of time. A party’s likelihood of success in policy or institutional debates depends partly on its influence within the legislature: its size, relative to other parties; whether there is a stable majority in the legislature or a shifting set of voting coalitions and whether other parties are willing to co-operate with it. This last factor will depend on whether potential legislative allies are also electoral rivals. In these calculations, separatists are no different from other politicians: all legislative actors make short term calculations about the current balance of parties and medium term calculations on how to position themselves for the next election.  However, separatists have a third factor to consider. The potential for separation itself will affect how much time and energy they expend on policy decisions affecting their region, on campaigning for the next election and on shaping their current institution. To sum up, I test the argument that separatists' behaviour will vary over time in response to three external factors: the potential for separation, electoral competition and their status in the legislature.   1.4 Testing the argument: cases and data The scope of this research covers separatist parties which compete in central parliamentary systems. Separatist parties are defined as any party which was established to campaign for self-government for 19 its region, where the party leadership has made a recent and definitive statement reaffirming the party’s commitment to such autonomy. Such a statement might be found in an election manifesto, a press release or a speech from the party leader. The precise constitutional form of that self-government may range from establishing a federal system in a unitary state to setting up an autonomous self-governing body for that particular region, to full separation and the establishment of a new country. The crucial theoretical distinction between what I term a “separatist” party and other parties is that the former's primary goal is new legislative and executive institutions for its region. Success for a separatist party therefore allows it to leave the institution to which it is currently elected: either because it has an alternative venue at the regional level or because it has seceded from the country altogether.  1.4.1 Country cases I have selected two parliamentary systems in which to test my arguments about separatist parties: the United Kingdom and Canada which occupy different positions in Lijphart’s typology of power-concentration in democratic systems (2012).  This typology consists of two dimensions: the “executive-parties” dimension, which measures the degree to which political authority is concentrated within central institutions and the “federal-unitary” dimension which measures the division of power between the central government and the regions. At the time of Lijphart's initial classification, Canada had dispersed power on the federal-unity dimension but had highly concentrated power on the executive-parties dimension. Until the late 1990s, the UK had highly concentrated political authority on both dimensions: with a centralized government which in turn was controlled by a small number of actors. In 1999, the UK devolved some legislative and executive powers to regional bodies, including the 20 Scottish Parliament, although it has retained its majoritarian electoral system and a party system based around two dominant parties.17   Both political systems, in all time periods, are therefore classed as majoritarian in at least one respect, erecting significant barriers to parties which are outside the political mainstream and which openly challenge existing constitutional and territorial arrangements. These countries are therefore hard cases, where we would not expect separatist parties to have significant influence within the political system. However, they also vary between themselves and over time in their degree of majoritarianism: they comprise unitary, partly devolved and federal structures, an essentially two-party system, and a multi-party system. If my arguments are supported by the evidence in these countries, therefore, they will not only be generalizable across a range of parliamentary systems but they will also give us an important insight into how separatist parties operate under inhospitable conditions.   1.4.2 Party cases Within these two countries, I focus on three separatist parties, selected because they have particular features of analytical interest.  Each party has participated in the central legislature for at least eighteen years, allowing us to track whether its behaviour varies over time, and if so in response to which conditions. These parties’ legislative behaviour will also be compared to the parties within their own political system. This allows me to test whether separatist parties behave differently to mainstream parties while holding constant factors such as constitutional structure, political culture, institutional rules and electoral and party system.                                                    17 The 2010-2015 coalition government of the Conservatives and Liberal Democrats does not change this analysis, as the electoral system remained the same, meaning that the inclusion of more than one party in government is likely to be an aberration.  21 The first case is the Bloc Québécois (BQ), formed in 1990 to seek the political and economic sovereignty of Québec. In the 1993 elections it became the second largest party in the Canadian Parliament and accepted the role of Official Opposition. Two years later, Québec’s provincial government held a referendum in which sovereignty was voted down, by a margin of just 1%. After this failure of its key demand, the Bloc continued to run for Parliament in the next six elections, winning 12-18 % of the seats in the House of Commons. It remaining a key player in Parliament until its collapse in the 2011 elections, when it was reduced to just 1% of the seats. It was the third largest party between 1997 and 2011, which gave it a pivotal position during the latter half of this period, when no party was able to win a majority of the seats in the House of Commons.   The case of the BQ allows a before and after comparison based around a crucial point in time: the 1995 referendum on Québec sovereignty. In the run-up to the referendum, all of Canada’s political actors knew that a yes vote and the subsequent break-up of the country was a distinct possibility, whereas after the results came in, separation was off the immediate political agenda. This allows us to test whether behaviour changes when the party’s expectations of the future change – from the immediacy of a referendum on secession to the medium term electoral cycle and continued participation in the Canadian Parliament over the next sixteen years. Studying the Bloc also allows other fruitful comparisons: between periods of majority government where it had no direct influence over legislative outcomes, and periods of minority government where it was pivotal.   The second case is the Scottish National Party (SNP). Prior to the 2015 UK election, the party had never held more that 1% of the seats at Westminster. It was first elected to the British Parliament in 1974, on a platform of independence for Scotland. Following the failure of the 1979 referendum on Scottish devolution, the SNP continued to run for statewide election and engage in parliamentary 22 activities during a long period when any degree of regional autonomy was off the political agenda. In 1998-9, under a new government, power was devolved to a newly established Scottish Parliament. However, the SNP chose to continue running in both British and Scottish parliamentary elections. The SNP has been in government at the sub-national level since 2008 and held an (unsuccessful) referendum on independence in September 2014. This echoes the 1995 Canadian episode: an actively separatist party engaging in the central legislature while the future shape of the country is decided.   Finally, a second party from the UK’s history is included: the Irish Parliamentary Party (IPP) which sought self-government for Ireland and sat at Westminster from the 1870s until it lost nearly all its seats to Sinn Féin in 1918.18  The Irish Nationalists are an example of a movement which actually did obtain legislation for self-government.19 They did so through both an alliance with one of the mainstream parties of the day (the Liberals) and the use of their pivotal status in Parliament to put pressure on governments from both parties. The IPP therefore allows us to examine the impact of legislative status on the behaviour of a separatist party. The party also went through periods when Home Rule was on the political agenda and periods when an Irish Parliament was a distant possibility. So, this case allows us to look at the effect that the prospects for self-government has on separatist campaign rhetoric.   1.4.3 Data sources The data used to test the argument about separatist parties focus on two types of activity: parties’ legislative behaviour, with data drawn from the official parliamentary record, and communications from politicians to voters during electoral campaigns. Sources for the latter type of data include                                                  18 The Irish Parliamentary Party went through several splits, reunions and consequent name changes during its history – see Lyons 1973.  I call it the IPP throughout, for ease of reference.  19 Although the Government of Ireland Act 1914 was never implemented due to World War I.  23 manifestos and campaign material produced by both the party’s central bureaucracy and its individual MPs. This should be material over which either the party’s campaign staff or an individual MP has had complete control: where their message is conveyed directly to their constituents, unmediated by the framing, opinions or rebuttals of others.   The key assumption underlying this choice of data is that legislative time and voter attention are both scarce resources. MPs have many demands on their time, so legislative participation must compete with activities with a clearer electoral or career pay-off such as constituency service, party work or direct election campaigns.  Although it is the government which controls the agenda for most parliamentary business, this still provides a menu of issue areas and types of legislative activity from which legislators and their party leadership can choose. Examining the issues on which separatists expend parliamentary time therefore gives us information about their relative preferences at particular points in time. Observing whether this varies with their legislative strength gives us information on the expected utility of these activities. This information therefore allows us to not only rank separatists’ preferences, but to infer which conditions they believe increase their chances of fulfilling those preferences.   How do we distinguish between issues that separatists care about for their own sake and those which help them win votes? When we observe separatists voting on, for example, fishing quotas, we cannot simply infer that they are doing so because they have a policy preference on the issue. Alternative explanations include a calculation that it will gain them votes at the next election, or a deal with another party which exchanges support on fishing quotas for concessions on self-government, another policy area entirely, or on procedural reforms within the central Parliament. The electoral explanation can be tested by examining the party’s own campaign material and individual legislators’ communications with their constituents. Campaign budgets are limited, so have to be targeted at 24 activities which will gain the largest return in votes.  Even communications with low marginal costs have to compete for attention against the messages of other parties, not to mention the myriad of other, non-political concerns of the average voter. Therefore, when a political actor purposely draws the public’s attention to her own parliamentary activity, this is good evidence that she believes it will win her votes.    Finally, comparing the same party’s legislative behaviour in periods when separation is on the immediate political agenda to times when it is not, can help distinguish between primary and secondary institutional preferences. When separation is not imminent, separatist legislators expect to remain in their current institution for some time. They therefore have incentives to spend time on parliamentary activities, for example to pursue procedural change which will benefit them in this legislature. Clearly, when separation is either a real possibility, or already decided, the high value they place on self-government should override any secondary concerns about the operation of an institution they plan to leave in the near future.  1.5 Organization of the dissertation This dissertation examines electoral and legislative behaviour to assess how separatists resolve the dilemmas they face when participating in the political system of the country they wish to leave. Chapter 2 sets out an overarching theory of separatist party behaviour, explaining how separatist parties differ from other types of party: in their preferences, their voter base and their access to resources. The chapter carries on to spell out how these differences lead to dilemmas in the electoral and legislative realms, before arguing which factors affect how separatist parties resolve these dilemmas. I conclude with hypotheses of separatist party behaviour under different electoral and legislative conditions, and when separation is on and off the agenda.  25 Part II (chapters 3 -5) tests these hypotheses against data on the legislative and campaign activities of the BQ, SNP and IPP. The empirical chapters of the dissertation use two levels of comparison through which to test my arguments. First, the length of these parties’ legislative records permits temporal comparisons, testing whether a separatist party’s behaviour changes in response to legislative and electoral conditions, and to the likelihood of separation. Second, in Chapter 5, I utilize within-country comparison to determine whether separatists' legislative behaviour is significantly different from mainstream parties and from small but pro-system parties.   Each chapter deals with a specific separatist dilemma. Chapter 3 compares election campaigns when self-government was imminent with campaigns when it was not, to assess the impact of time horizons on separatist party rhetoric. Chapter 4 also focuses on campaign behaviour, but focuses on elections when separation is off the political agenda. It asks when separatists will raise grievances and when they will cooperate with the governing party, in order to claim credit for benefits their region has received. Chapter 5 focuses on separatists' level of participation in the legislature: do they seek to change the procedural disadvantages in the central parliament and will they invest time engaging with an institution they wish to leave?   Chapter 6 draws the results of the empirical analysis together to assess the success of my theory of separatist legislative behaviour and to establish the remaining gaps in our knowledge.  It also explains how this research furthers our understanding of legislatures and parties, while making a contribution to broader institutional theory: showing how the potential for exit affects institutional design and maintenance and interactions with other actors. It closes by considering the normative implications of my findings: are developed democracies sufficiently robust to include  actors that challenge the whole system? 26 Chapter 2:  Theoretical framework This chapter sets out the theoretical argument about separatist party behaviour that is tested in Part II of the dissertation. I begin with a simplified model of a party system, comparing separatist parties to the mainstream government and opposition parties. I argue that separatists face challenges in both the electoral and legislative realms: they have limited legislative resources but must appeal to an ideologically diverse group of voters while trying to achieve a long-term aim which is opposed by the other parties. The second section explains how these constraints lead to dilemmas which are quite distinct from those of the other parties, affecting separatists' electoral and legislative behaviour. The third section then details how we would expect separatists to respond to these dilemmas. I argue that the party leadership’s priorities will shift in response to the potential for separation. Equally, opportunities to pursue a separatist agenda will vary with the party’s status in the legislature and with the electoral competition it faces. The chapter then concludes with hypotheses of separatist party behaviour over time.    2.1 Separatists in a central party system I begin by positioning separatists within a classic majoritarian system, where single-member districts and plurality electoral rules encourage candidates to consolidate into a two party system (see, e.g. Cox, 1997; Duverger, 1954; Lijphart, 2012). These two mainstream parties alternate between government and opposition.  In such a system, small parties normally find it difficult to build enough support to win seats (Ezrow, 2010). However this effect only applies to parties with wide but shallow support spread across the entire electorate. Electoral support for separatists is necessarily territorially concentrated. This means their support is divided across fewer constituency boundaries, making it easier for 27 separatists than for other small parties to win parliamentary seats.20  In addition, campaign resources are not spread as thinly as they are by small parties that are running countrywide. For these reasons separatists find it easier to get into the legislature than other types of small party. This leaves us with a model in which two countrywide parties compete for government, plus a separatist party that only runs in its own region. How, then, does this separatist party differ from the mainstream parties?  2.1.1 Procedural focus As Meguid notes, at elections mainstream parties generally position themselves along a left-right dimension which covers a range of issues but can be summarized as the degree of government intervention in the economy and society (2008, 3-4). These parties appeal to the voters who are closest to them and furthest away from their competitors on this spectrum. This means that the voters for each party form a coherent group with similar ideological positions. When voting in the legislature, within a majoritarian parliamentary system, these mainstream parties tend to form into “government v opposition” blocks (Hix and Noury, 2014). As the government party wishes to stay in office and the opposition party seeks to replace it, an adversarial relationship develops. The government party defends its policy record and legislative agenda, while the opposition party attacks this record in order to persuade voters that it is a more worthy candidate for government.   Separatists, however, do not fit neatly into the government-opposition dichotomy. They neither lead the central government nor make a plausible government in waiting: as separatists have no appeal outside their own region, they have virtually no chance of growing to become a mainstream party within the current political system. In terms of policy, separatists have concluded that their region’s needs cannot be met within the current state and therefore seek a new set of constitutional and territorial                                                  20 See Meguid, 2008, 7 for a review of the literature on this point. 28 arrangements. The separatist leadership’s primary concern is therefore not to get into government but to leave the system, meaning that it has to turn voters against the whole regime, not simply the government of the day. Such an aim does not translate into a specific position on the left-right policy scale. For separatists then, the debate is not primarily about which group or which policy objective gets the most resources, nor the government’s role in providing those resources. Rather, it is a procedural question of who makes such decisions on behalf of whom in the first place.   2.1.2 Ideologically diverse vote base Separatists can therefore link any issue to their primary concern, by stating that decisions affecting their constituents should simply not be made by the current institution. This means that separatists have a procedural argument to make on a very wide range of issues, allowing them to engage voters with differing policy interests. At the same time, separatist leaders can downplay their substantive position on any of these policy areas, creating the opportunity to draw voters from across the left-right political spectrum.   Naturally, the core of these voters are separatists themselves. However, the party’s regional focus allows it to compete for two other groups of voters: those seeking greater resources for their region (benefit-seeking voters) and those that wish to protest against the incumbent government by supporting an opposition party (protest voters). Separatist parties appeal to benefit-seekers and protest voters with rhetoric based on identity rather than policy positions. By definition they represent only their own ethno-territorial group so do not have to accommodate the wishes of those outside that group. Meanwhile mainstream parties are campaigning across the state, so cannot openly favour one region over another.  In this way regional parties drive a different form of political conflict than that on the left-right spectrum. Rather than focusing on socio-economic divisions, the debate pits regions against 29 one another, encouraging voters to think about questions of distribution in regional (instead of class) terms.21 Benefit-seeking and protest voters are therefore given a rallying point around which to make demands and express grievances.   So, separatist parties have the opportunity to attract voters from a range of socio-economic backgrounds, and from across the right-left spectrum. Furthermore, they have an advantage over other small parties in converting that support into actual seats in the legislature. However, once in Parliament, they need to demonstrate to their electoral coalition that they can either obtain benefits for the region, attack the government or press for constitutional reform. To do so, they need to have an impact on legislative proceedings.  2.1.3 Limited resources Regional concentration helps a separatist party get into Parliament but puts a fixed ceiling on its growth. These limited caucus numbers then disadvantage separatists in day to day legislative activities. For any Member of Parliament, attendance carries opportunity costs: it is time not spent in the constituency, either building up a personal following (Kam, 2009) or taking the party’s message directly to voters. However, unlike government MPs, separatists cannot compensate for lost constituency time with increased policy influence, as they don’t have enough members to win votes on the issues they care about. When the government of the day is stable, small parties cannot even influence the outcome of votes through alliances with another party, leaving them very limited leverage with which to demand concessions.                                                   21 See Detterbeck and Hepburn, 2010 and Hepburn, 2011 for an examination of the territorial dimension of competition in European party systems and Cairns, 1968; Godbout and Høyland, 2011 and Johnston, 2005 on territorialism in the Canadian context. 30 Outside the main chamber, the separatist party’s size means that it is only granted a limited number of seats on committees. It therefore doesn’t have the same opportunities as the mainstream opposition party to scrutinize government policy and legislation. In addition to this lack of policy influence, members of small parties are handicapped in building their profile with the electorate. Legislative time (whether it is used to debate, ask questions or propose motions) is a scarce resource and will be disproportionately devoted to the business of the governing party, followed by the official opposition (Cox, 2005). This leaves separatists with only limited chances to advance their case.   This situation is exacerbated by the anti-system stance of a separatist party. There is no reason for pro-system, statewide parties to want to grant parliamentary time to attacks on the overall political system or to the provocation of inter-regional conflict. Successive governments (whatever their position on the left-right spectrum) therefore have both the means and the motive to deny separatists a platform, leaving the latter with little material on which to campaign at the next election. This makes it difficult for separatists to show voters that sending the party to the central legislature has resulted in either policy change or many opportunities to advocate the region’s needs to a countrywide audience.   These procedural handicaps facing separatists are compounded by the impact of parliamentary time on media exposure. Countrywide media outlets are in any case unlikely to cover parties which have limited influence in the legislature and for which only a portion of the audience can vote. Regional media, though, do have an incentive to cover parties whose support is concentrated in their part of the country. However, in order to keep the audience’s attention, coverage of legislative proceedings is likely to focus on high profile or acrimonious debates and question periods. If separatists are unable to make many contributions on such occasions, there will be limited material with which to illustrate their parliamentary activities. In this case, voters may see regional media stories about separatists’ ideology 31 or local activities, but cannot be shown what such politicians do when they are elected to the central legislature.   The net effect of these differences between separatists and mainstream parties is that the former have to appeal to a wide range of voters with only limited resources. They do so while grappling with the paradox of engaging with the very political system they reject. This creates a series of dilemmas for the party leadership, affecting how it campaigns for re-election and whether it participates in day to day legislative activities.   2.2 The dilemmas of a separatist party As discussed in Chapter 1, mainstream legislators have to juggle the pursuit of government office and electoral concerns with their policy preferences. They pursue these goals through parliamentary activities, such as debates, proposing bills and motions, committee work etc. These activities allow them to implement their agenda when in government, to scrutinize legislation and government activities when in opposition and to build a public profile on which they can campaign at the next election. Do separatist MPs have the same concerns?  With only a regional base, the prospect of separatists getting into central government is remote. In any case, taking office can undermine the party’s anti-system stance. However, winning more seats is still a worthwhile goal for the party leadership. It increases the chances of being pivotal in future parliaments with the accompanying chance to win concessions. Participation in parliamentary activities therefore serves two functions for separatists. First, advocating their primary preference for self-government and second, building their profile with the electorate which serves the aim of electoral growth.    32 These two goals mean that separatists cannot ignore the organization of Parliament itself.  The procedural rules which determine speaking priorities, committee appointments, access to research resources and staffing budgets will all affect the capacity of legislators to pursue their aims. Separatist MPs have chosen to pursue their goals within the existing political system, so they benefit from sitting on the committees which govern the day-to-day operation of the legislature. Equally, if they participate in debates on the reform of Parliament’s practices and help to develop the institution’s rules, regulations and norms, they can attempt to address some of the procedural hardships facing small parties.   So, like mainstream MPs, separatists have policy, electoral and institutional concerns. Their secondary goals of electoral growth and reform to the central legislature can be achieved within the current political system. However, separatists differ from other parties in their primary objective. They wish to leave the political system and this outweighs the value of securing particular policy objectives or government office within it. Moreover, separatist parties are trying to achieve this with only limited caucus numbers and restricted parliamentary time at their disposal. This means that the leadership has to prioritize how it uses these resources to meet its region’s needs. The key choice the leadership faces is how best to meet those needs: by preparing for an independent political system, or by advocating for the region within the institution where the party currently holds seats? This section sets out how this “future system versus current system” dilemma plays out for a separatist party, in both its electoral and legislative activities.  2.2.1 The choices open to separatist parties: electoral strategy Turning first to the goal of electoral growth, the party can appeal to a diverse range of voters, across the left-right spectrum. However, it has to emphasize regional identity and explain why it is better to send 33 candidates from a regional, anti-system party to Parliament, as opposed to candidates from the countrywide mainstream parties.  To do so, separatists can choose between the electoral strategies used by the mainstream parties: either grieving about problems, as the opposition party does or claiming credit for popular programs, as the governing party does. The former strategy seeks to convince voters that the status quo does not serve their needs, thus laying the groundwork for public acceptance of a new, self-governing political system for the region. Conversely, a credit-claiming strategy argues that it was separatists who pressured the government to provide benefits to the region. This approach reinforces the role that separatists have in the current political system, as a regional champion.    Separatists differ from the mainstream parties in actually having a choice. The governing party has to defend its own record and is obviously best placed to claim the credit for a popular program as it controls the passage of legislation and the allocation of public expenditure. Conversely, the mainstream opposition party has to attack the government’s activities. This draws attention to the negative impact of government policies and to those groups which have not received benefits: in other words, a grievance strategy. The opposition is seeking office itself and therefore has to undermine the incumbent government’s appeal to voters. Separatists, meanwhile, are not plausible candidates for government and do not have to balance different demands from the various parts of the country. They can therefore position themselves as their region’s advocate in the central legislature: there to obtain benefits and to raise the region’s historical, cultural or economic grievances.  Opposition parties could attempt a credit-claiming appeal to benefit-seeking voters by arguing that their MPs put pressure on ministers to make a particular policy or budgetary decision. However, such an argument is only credible if benefits are actually delivered to the voters, which ultimately rests on the governing party’s decision. A mainstream opposition party is therefore unlikely to credit-claim, as this 34 implicitly endorses the actions of the very government it seeks to replace. Such an approach is open to a party that does not aspire to central government. Indeed, when using the credit-claiming strategy, separatists argue that it is their party which is best placed to represent the region, whichever mainstream party is in government.    The alternative campaign strategy is to appeal to protest voters by attacking the current government. However, separatist politicians also have to maintain their appeal to separatist voters and convince benefit-seeking and protest voters of the case for secession. Separatist grieving may therefore differ from that of mainstream opposition parties. The latter want to attack the record of the current government, suggesting that voters can achieve worthwhile change simply by electing a different party  to office. Separatists, meanwhile, need to show that their region’s complaints are intrinsic to the current territorial and constitutional set-up and cannot be solved by simply changing governments. The grievances raised by separatist politicians are therefore likely to attack the entire regime, not just the government of the day.   There are electoral trade-offs in the choice between credit-claiming and grieving. Grieving appeals to protest voters and genuinely separatist voters, but if it attacks the political system itself, it may be too extreme for voters who simply want more benefits for their region.  A credit-claiming strategy, meanwhile, placates benefit-seeking voters but at the cost of having to share that credit with the government. An anti-system party which does this can be accused of hypocrisy for co-operating with the political establishment, thus alienating voters who are committed to separatism.   The trade-off discussed above is between different groups of voters. However, a more serious problem is that election rhetoric can force separatists to choose between their primary goal of self-government 35 and the secondary goal of electoral growth. Using credit-claiming may win the party more votes, but it does so by pointing to benefits that the region receives under the current political system. This strategy therefore risks undermining the case for separatism. Constituents may vote for the party as an effective provider for the region, but such electoral support will not translate into winning a future referendum on separation, nor into public support for constitutional negotiations with the government, thus risking the primary aim of the party. Therefore, in designing its electoral strategy, the party must choose between its role as regional advocate within the central legislature, where power currently resides, and the advocate for separation and an entirely new system of governance.  2.2.2 The choices open to separatist parties: institutional participation If the electoral campaign is successful, separatists then have to decide how to use their seats in Parliament. This raises the second dilemma facing the separatist leadership. Politicians’ capacity to affect policy and to scrutinize their opponents is affected by the rules and procedures of the legislature. Should separatists therefore participate in the day to day operation of the central legislature and discussions of reform for the current political system? Or should they concentrate instead on the new political system they seek?  Focusing on a new system of governance for their region reflects separatists’ primary preferences and will also appeal to their core separatist voters. There are however, two possible approaches to focusing on the future, which lead to very different sets of behaviour. The party could use the central legislature as a platform to gain publicity for separatism and as a forum in which to gain concessions from other political actors on self-government. They can achieve this directly by engaging in debates about the terms of separation and the jurisdiction and powers of any new institutions. Alternatively the party leadership could decide to disengage from the central legislature, underlining its rejection of the 36 constitutional status quo. In this case, separatists would run for election in order to emphasize the degree of separatist sentiment among the region’s citizens: the campaign would be their platform rather than the legislature. Separatist MPs would then focus on constituency work and would use regional media to advance the case for self-government. However, they would not engage in debate within the legislature itself.   If the party instead focused on its secondary goal of electoral growth, it would be less concerned with the powers and design of future, regional institutions and more focused on the operation of the current, state-wide legislature. We would expect to see more participation in day to day parliamentary activities and more concern about the party's status within the institution. In particular, separatist MPs should participate in reforms to the state-wide legislature, seeking to reduce the procedural disadvantages faced by small parties such as themselves. Such reforms would be valuable because they would help the separatist party access greater amounts of parliamentary time. Gaining procedural advantages in the legislature could be used to advance the case for separation itself. However, if the party is concerned about electoral performance, self-government will not be the only focus. They should also use the legislature to seek benefits such as additional investment in their region. The extra parliamentary time and additional resources could therefore help address the region’s concerns and build a platform for the next election.  As with electoral choices, the separatist leadership must make trade-offs in deciding its level of participation. If the party’s MPs engage with central institutions they dilute their own message that the constitutional status quo is unacceptable. This risks losing credibility with their core voters, who are also opposed to the current system.  The leadership must therefore decide whether to invest time and resources in participating in an institution from which it may not benefit over the long term.  However, 37 boycotting such activities will add further to the party’s procedural disadvantage in the legislature, denying them a platform from which to win votes. Such disengagement may also alienate protest or benefit-focused voters who are not fundamentally opposed to the system as it stands. It is difficult to explain to such voters why they should use their votes in a countrywide election to support a party which will not actually participate in the central political process.   Table 2.1 summarizes the central dilemma of separatists: whether to focus on the future political system they want or the current system they are in.  The figure breaks this dilemma down into the electoral and legislative realms – these form the dependent variables in this research project.   Table 2.1 The electoral and legislative choices of a separatist party  Focusing on future political system Focusing on current political system  Electoral strategy  Grievances against political system.  OR Making the case for self-government   Credit-claiming.   OR Grievances against government of the day   Legislative strategy  Seek constitutional reform: self-government OR Disengage from central legislature  Seek increased government services for their region OR Seek reform of central legislature: increase status of small parties  The following section sets out the independent variables which determine the choices that separatist leaders will actually make, and under which circumstances.  38 2.3 Resolving separatist dilemmas: constitutional, electoral and legislative opportunities  So, in both electoral and legislative activities, separatists face a dilemma over whether to focus on a future, independent political system or on the existing political system in which they actually operate. Either approach involves trade-offs between different parts of the separatists’ voter base. It also requires prioritizing between the goal of separation and the party’s status within the current system. This section explains how separatists resolve these trade-offs, explaining how their priorities change over time in response to environmental factors.  2.3.1 The potential for separation The most important factor driving separatists’ choices is whether their primary objective of self-government is on the political agenda or not.22  When this kind of change is on the agenda, we would expect all of the separatists’ activities to be focused on the new political system they seek, not the current central institutions. Their electoral campaigns are therefore likely to favour anti-system grieving, with separatists spelling out what is wrong with the existing system from their region’s perspective and offering a solution: self-government. Similarly, we would not expect separatists to prioritize the institutional design of a legislature they may soon leave, nor regional investment from a government they repudiate. Rather, they should focus on the powers to be given to the new set of political institutions, and on winning any referendum on self-government.  For separatists, grieving against the political system helps win them votes from constituents who are already committed to separatism. However, use of this rhetoric is not only used to win these votes. It also helps maximize the public appeal of leaving the country, which is particularly important if the question is being put to a referendum. This will trump all lower priorities such as winning seats in the                                                  22 The possibility of exit may be due to a referendum on separation itself or as part of a package of wider constitutional reforms. 39 central legislature.  The separatist leadership will therefore not wish to draw attention to the benefits the region derives from the status quo - this rules out a pro-system, credit-claiming electoral strategy. Conversely, when separation is off the agenda, such as after a no vote in a referendum, the separatist party is under pressure to show that it is still relevant within the current system. It can win votes as a champion for the region, as a medium term strategy to stay in the legislature until separation returns to the agenda. The potential for separation, therefore, affects the separatist party’s choice of one campaign strategy over the other: when self-government is on the agenda we should see an anti-system, grievance-based campaign, whereas when it is off the agenda, a credit-claiming strategy is more likely.                                                    When there is no immediate prospect for self-government, separatists have to continue to participate in the central political system for the foreseeable future. This gives them the incentive to get involved in the day to day operation of the central legislature and to support reforms of the current institution which would increase their own influence at the statewide level. We would also expect them to use legislative time to demand benefits for their region (such as increased government spending), in order to prepare the way for credit-claiming rhetoric in the next election.  2.3.2 Electoral competition with the governing party                             The primary objective of leaving the political system is served by the secondary goal of winning seats in the central legislature. The latter allows separatists to demonstrate the strength of secessionist sentiment among the public, putting pressure on the government of the day to make concessions. The separatist leadership should therefore devote time to planning its next electoral campaign and determining which parliamentary seats it has the best chance of winning.       40 If separatists are primarily competing with the government party23 for these seats, then neither side has any motivation to share the credit for benefits given to the region. Doing so would simply risk losing benefit-seeking voters to their opponents. The separatists’ best strategy in this circumstance is to concentrate on appeasing its core separatist vote and attracting protest voters by laying out the region’s grievances and attacking the government’s record in addressing those problems. Grieving serves both the primary, long-term goal of persuading a majority of the region’s citizens to support separation and the secondary goal of depressing the government’s vote in the region, thus helping the separatists to win more seats.   The above sections explain separatists’ incentives to use a particular campaign strategy or to seek concessions on institutional design. However, some of these choices require the co-operation of other parties. A separatist party can only claim the credit for public investment if the region actually received these benefits, which is ultimately a government decision.  Equally, separatists will only be able to win concessions on self-government or influence the day to day management of the central legislature if other parties include them in such decisions. Where this is not the case, we would be more likely to see the party disengage from the central legislature altogether. Such a course of action would underline the party’s rejection of the current state and play to its support base of separatist voters. It is therefore necessary to consider when separatists will have the opportunity to put their preferences into action at the statewide level.  2.3.3 Electoral competition with the opposition party As explained previously, when the separatist party is competing with the government party for seats, neither side has any motivation to share the credit for policy benefits given to the region. If, however,                                                  23 The section below deals with cases where the separatist party is not competing for seats against the government. 41 separatists are in tight races with the opposition party, the incentives are reversed, creating an opportunity for separatists to use the credit-claiming electoral strategy. From the government’s perspective, separatists are not a threat on the central stage. As they only run regionally, they cannot win a majority of the legislature’s seats and are therefore not an alternative government. If the region is one where the opposition party usually does well, contests won by separatists reduce the opposition’s overall tally of seats, making it easier for the governing party to retain a plurality of the seats in the whole legislature. The government therefore has a motive to shore up the separatists’ vote and can do this by sending benefits to the region, arguing that this is in response to pressure from the separatist party. This is a particularly attractive strategy as it allows the government to weaken the mainstream opposition without having to expend scarce campaign resources in electoral districts where it is not competitive.  The opposition party will not want to engage in any activity which praises the incumbent government, for fear of undermining its statewide campaign. It therefore won’t be able to make an appeal to the region’s benefit-seeking voters, giving the separatists a clear run at this section of the electorate.  In such circumstances, the anti-government section of the electorate (protest voters) will find a natural home with the opposition party, giving a further reason for the separatists to concentrate on benefit-seeking voters instead.   2.3.4 Pivotal status If the government does not have a working majority of the seats in Parliament, this creates an opportunity for separatists to seek concessions. The separatist party may achieve these by joining the government in return for a pre-agreed package of measures. Alternatively they can offer ad hoc support on matters of confidence and supply, in return for specific policy compromises, as and when these 42 issues come before Parliament. A third option is combining with other opposition parties to defeat the government. Whichever route they take, separatists can use this leverage to seek distributive, institutional or constitutional concessions. The possibilities include benefits for their region, legislative time, seats on committees, greater autonomy for regional bodies or a referendum on separation. In terms of electoral strategy, pivotal status makes it more likely that a separatist party will use credit-claiming rather than grieving. Such leverage gives the separatists far more credibility when claiming that they were the cause of specific budgetary decisions.    The effect of pivotal status on the separatists’ legislative priorities is not so clear-cut: it gives them the opportunity to seek concessions focused either on a new constitutional structure, or within the current set of institutions. Their choice will depend on the factors outlined in the previous section. If separation is off the agenda, separatists will be operating within the current system for the foreseeable future. Their priority would therefore be to ensure they can function within the central legislature. They would want their MPs to be able to sit on committees and scrutinize the impact of legislation and policy on their region (thus giving them material which can be used for either a grieving or credit-claiming strategy in future elections). The party leadership would also be concerned with its share of legislative time: if it is going to use the legislature as a platform for espousing the separatist cause and for building up its profile with the electorate, its members need to participate in high profile debates and be called to ask questions of the government. In circumstances where separatists are pivotal but separation is not on the agenda, we would therefore expect them to participate in initiatives to reform the central legislature, in order to protect the interests of small parties. Similarly, they should engage in day to day decisions about the operation of the legislature such as the scheduling of votes and the allocation of legislative time. As these activities involve a degree of cross-party discussion and compromise, we should see separatists co-operating with both of the mainstream parties in order to gain institutional concessions.  43 Conversely, separatists should use their pivotal status in a different way when separation is on the political agenda. Here, we would expect them to focus on the future: on the autonomy of their region. In these circumstances, separatists have compelling reasons to co-operate with the government rather than the opposition party, in order to influence the design of the new political system that would follow self-government. Even as a minority government it still has agenda control in Parliament and the machinery of the state behind it. The government decides which issue areas get the most time in the legislature and which policy problems are prioritized by public servants. We should therefore see an increase in separatists' legislative engagement at times when they are pivotal. If self-government is on the agenda at the same time, that engagement should primarily take the form of co-operation with governing parties.                                                                                                                                                                Figure 2.1, below, summarizes how the separatist parties’ opportunities to gain seats and to gain concessions will affect their electoral and legislative activities. The y-axis represents whether separation is on the agenda or not. When it is, separatists will prioritize their primary goal of self-government and their behaviour will be driven by preparation for a new political system. When it is not, they have to focus on the political system they are currently in, so they prioritize secondary goals of electoral growth, benefits for the region and procedural reforms to aid small parties. The x-axis represents the separatist party’s opportunities to pursue its goals. Such opportunities arise when the party is pivotal in the legislature, or when it competes electorally with the mainstream opposition party. In both cases, the government party is motivated to grant concessions to the separatists. The content of the cells spells out the behaviour we expect in each situation: when primary goals are the priority with opportunities to gain concessions (upper left quadrant) versus no such opportunities (upper right quadrant) and when secondary goals are the priority with opportunities for concessions (lower left quadrant) versus no opportunities (lower right quadrant). 44 2.4 Interaction of the independent variables So, when selecting an electoral strategy, we would expect a separatist leadership to consider the potential for separation, the party's electoral competitor and the opportunities provided by its status in the legislature. Likewise, when choosing whether (and how) to engage in the legislature, these same leaders will consider the likelihood of a new political system being established and their status in the current one. Of course, in the political systems we observe, these factors do not occur in isolation. Table 2.2, below, sets out all the possible combinations of the three independent variables. Considering how these factors interact helps us to understand which have a greater impact on separatist behaviour than others.   45  Figure 2.1 Separatist behavior: constitutional, electoral and legislative opportunities     Separation on the agenda: pursuing primary goals  Electoral strategy:  grieving against the political system and making positive case for separation  Legislative behaviour: negotiating self-government  Electoral strategy: grieving against the political system and making positive case for separation    Legislative behaviour: disengaging from central parliament  Electoral strategy: credit-claiming   Legislative behaviour:     - seek increased time/resources for small parties in central legislature   - seek policy benefits for region  Electoral strategy: grieving against the government of the day  Legislative behaviour:   use as platform to build public profile of region Separation not on the agenda: pursuing secondary goals  Opportunities for concessions      No opportunities for concessions  46  Table 2.2 Separatist campaign behaviour: the interacting effects of (1) prospects for self-government,  (2) pivotal status and (3) electoral competitor  Self-government on political agenda: short-term focus on primary goal Self-government off political agenda: long-term focus on primary goal; medium-term focus on electoral gains  Separatists pivotal: opportunity to credit-claim Separatists not pivotal: no opportunity to credit-claim Separatists pivotal: opportunity to credit-claim Separatists not pivotal: no opportunity to credit-claim Government party is electoral competitor: medium-term electoral incentive to grieve; no opportunity to credit-claim Scenario I:   Grieve against the political system  Scenario III:  Grieve against the political system         Scenario V:  Grieve against the government (despite Gvt being willing to share credit) Scenario VII:  Grieve against the government Opposition party is electoral competitor: opportunity to credit-claim Scenario II:  Grieve against the political system  Scenario IV:  Grieve against the political system  Scenario VI:  Credit-claim Scenario VIII:  Credit-claim (opportunity arises due to electoral competition with opposition)   2.4.1 The impact on campaign strategy. The first factor, potential for separation, concerns the party's primary preferences. When self-government is on the agenda, the separatists’ campaign strategy can affect the outcome of a referendum as well as statewide elections.  By contrast, pivotal status and the identity of the electoral competitor affect opportunities to credit-claim and hence the secondary desire to win more seats at the next election. So, we would expect the potential for separation to trump all other factors. For example, if a referendum on separation was scheduled, but the separatist party was pivotal and its main electoral rival was an opposition party (scenario II in Table 2.2 above), we would still expect a grievance 47 strategy to dominate, even though the latter two factors favour a credit-claiming strategy. This is because the primary objective of self-government will outweigh the utility of electoral advantage in a system which the party will soon have the chance to exit. The presence of separation on the political agenda is therefore expected to lead to grieving, regardless of the separatists’ legislative position, or the shape of electoral competition (see scenarios I-IV in Table 2.2 above).    What about the interaction between legislative status and electoral competitor? There are two situations which create conflicting pressures, assuming that separation is not on the political agenda. First, a pivotal separatist party which directly competes with the government for seats has a legislative opportunity to gain concessions, and therefore to credit-claim but an electoral reason to grieve (scenario V in Table 2.2). Second, a non-pivotal separatist party which competes with the main opposition party has limited legislative leverage with which to demand policy benefits and therefore no opportunity to credit-claim. However, it has an electoral incentive to do so (scenario VIII in Table 2.2 above).24   In each case, we should first consider the governing party’s incentives, as credit-claiming cannot occur without it deciding to provide benefits to the region in the first place. In the first case (scenario V), we would expect the government’s short term desire to remain in office to override its medium term concerns about losing seats to the separatists. So, if the separatists are pivotal, the government should be willing to send benefits to their region, and to share credit with the separatist party, in return for support on matters of confidence and supply. If there is no such deal and the government falls, its medium term electoral concerns would likely become short term. However, the government would then                                                  24 In the remaining two scenarios shown in Table 2.2, the two factors create reinforcing incentives. In scenario VI, both the separatists’ pivotal status and the fact that they compete with the main opposition party give the government a motive to share credit with them. This creates an opportunity for the separatists to credit-claim. In scenario VII, the party is not pivotal so has no opportunity to credit-claim, and competes with the governing party, so has an electoral incentive to grieve.  48 be going into an election having not given any benefits to the region in question, so would still be at a disadvantage in tight races. To summarize then, the government’s best strategy is to offer benefits to the region and share the credit with the separatists.   In the second case (scenario VIII), where separatists are competing with the opposition party and do not hold the legislative balance of power, the government is under no such immediate pressure to offer this kind of benefits and credit-sharing deal. However, the government can give itself a future electoral advantage by doing so. A further advantage from the government’s perspective is that supplying such benefits undermines the case for separation as it demonstrates what the existing state can offer.  Assuming that there are no countervailing short term concerns (such as demands from another party which is pivotal) it would be in the government’s interest to do so. In both cases, therefore, we should expect the government to offer to send benefits to the separatists’ home region and to share the credit for doing so – but will the separatist leadership accept?  As already explained, when a separatist party competes with the opposition party for seats, the latter is likely to appeal to protest voters. This encourages the separatist party to target benefit-seeking voters instead and the government’s actions provide them with the opportunity to do so.  In the second scenario, therefore, separatists would credit-claim. However, in the first scenario, where the separatist party’s direct competitor is the government itself, any electoral advantage gained from credit-claiming is offset by two risks. First, benefit-seeking voters may simply credit the government itself and cast their votes accordingly, against the separatists. Second, such a deal will simultaneously undermine the separatist party’s appeal to protest voters and its core supporters who are primarily concerned with exit. Therefore we expect separatists in this situation to concentrate on a grievance strategy. Rather than 49 using their leverage in the legislature to demand policy benefits for their region, they can instead use it to gain institutional concessions, which will be less visible to the electorate.   These interactive effects are summarized in Table 2.2, above, which shows the eight possible scenarios which result when the three independent variables occur together. The column and row headings set out how each independent variable affects the separatists’ opportunities, ceteris paribus. The numbered cells show the outcome when all three independent variables interact. For example, in scenario VIII, separatists are not pivotal, which taken by itself means they have no opportunity to extract benefits (and thereby credit-claim) by putting legislative pressure on the government. However, in this scenario, separatists also compete electorally with the opposition party, giving the government a different, electoral motive to share credit with the separatists. This creates a separate opportunity for the separatists to credit-claim.  2.4.2 The impact on legislative participation Here, the relevant factors are whether self-government is on the agenda and pivotal status. Table 2.3 sets out the four possible combinations of the independent variables, with the legislative behaviour we expect to see from a separatist party in each scenario. As with campaign strategy, the potential for self-government is the most important variable explaining separatists’ behaviour in the legislature. When separation is on the agenda, separatists will focus on this future political system, regardless of their status in the existing legislature. However, the latter factor does affect the tactics they use.  When pivotal, (scenario A in Table 2.3) separatist parties have an opportunity to seek further concessions, which they can use to safeguard the autonomy of their region’s future government. These concessions could be constitutional - on the design and jurisdiction of the region’s new institutions, or 50 could concern the process of holding a referendums on self-government. In order to achieve these objectives, the separatist party needs to participate in relevant debates within the legislature. However, if it is not pivotal, then we would expect future-focused separatists to simply disengage from the central legislature (scenario B in Table 2.3). Given that they have no opportunity to extract concessions and do not expect to remain in the current institution, participation detracts from time they could be spending in the region, convincing their constituents of the case for self-government.   When separation is off the agenda, the party will more likely focus on achieving its objectives within the current political system (see scenarios C and D in Table 2.3). The leadership expects to remain in the central political system for the foreseeable future, so has to focus on how to win seats in future elections. In this case, it needs to reach beyond its separatist support base and appeal to pro-system voters by painting itself as a champion for the region. It can do this using the central legislature as a platform. This means the party participates in legislative activities in order to advance the region’s concerns and win policy benefits for the region. In doing so it can build the party’s public profile. As there is no referendum on self-government in the near future, there is less danger that the separatists’ participation in the central legislature undermines the case for future separation.    However, its choice of activities will hinge on whether it is has leverage. When it is pivotal (scenario C in Table 2.3, below), the separatist leadership has the opportunity to overcome the procedural handicaps which usually face small parties. It therefore has an incentive to participate in day-to-day operational decisions such as the division of parliamentary time, in order to ensure that its region’s concerns are included on the agenda. Similarly, if there are any initiatives to reform the central legislature itself, or if the institution’s own rules and regulations are up for debate, then separatists can use such opportunities to increase their own status. They could, for example, seek more committee 51 placements, speaking opportunities or resources for their own Members. Conversely, when the separatist party is not pivotal, it has no such leverage on operational matters or questions of reform (scenario D in Table 2.3, below). Given the risk of alienating its own core voters, it would not engage in activities which are unlikely to bring it concessions in any case.   Table 2.3 Legislative participation: the interacting effects of (1) prospects for self-government and  (2) pivotal status  Self-government on political agenda:   short-term focus on new political system (primary goal) Self-government off political agenda:  long-term focus on new political system (primary goal); medium-term focus on electoral growth in current system (secondary goal) Separatists pivotal:   opportunity to gain concessions Scenario A  Focus on future political system:  engage in legislative debates on self-government  Scenario C  Focus on current political system:      use legislature as platform to                   build public profile  AND  - engage in reform initiatives for central legislature - participate in day to day operation of legislature - seek policy benefits for region Separatists not pivotal:  no opportunity to gain concessions Scenario B  Focus on future political system:   disengage from central legislature Scenario D  Focus on current political system:  use legislature as platform to build public profile  The argument laid out in this section explains how separatists’ opportunities and priorities shift in response to the potential for self-government, their status in the legislature and electoral competition. I have used these factors to predict how separatists’ campaign strategies and participation in legislative 52 activities will change over time.  The final section of this chapter summarizes this model in a series of hypotheses which will be tested in Part II of this dissertation.  2.4.3 Threats to causal inference Before moving on to the overall hypotheses, it is helpful to set out potential pitfalls in assessing the determinants of separatists’ behaviour. First, causation may run in the opposite direction to my argument. The separatist party’s own behaviour may be what actually puts self-government on the agenda. I rule out this possibility by selecting cases where a referendum on self-government was initiated by either the central or the sub-national government, so it is exogenous to the choices made by the separatist party caucus in the central legislature.25  Second, there is the possibility of spurious correlation. Government actions, changes in the overall party system, or other, un-theorized, variables could affect both the likelihood of self-government and the rhetoric used by separatists. Indeed, as noted above, for some of my cases it was the central government that initiated constitutional change. However, if government action was driving separatist behaviour, we would expect separatists to be supportive of the governments which granted more autonomy to the regions. This runs counter to my prediction that separatists will use a grievance strategy when self-government is on the agenda. When there has been a shake-up of the party system, separatists may find that their primary electoral competitor has changed. If this was indeed driving separatist behaviour, it would mean the separatists were more concerned with their electoral position in                                                  25 So, the 1995 Québec and 2014 Scottish referendums were introduced when a separatist party gained a majority in the sub-national legislatures (the latter after negotiation with the central government; BBC, 2012). The 1979 and 1997 Scottish referendums on devolution were the result of a change in central government, from a mainstream party that opposed devolution to one that supported it (Mitchell, 2009). The IPP is a more difficult case, as its legislative pressure on minority Liberal governments was at least partly responsible for the introduction if the Home Rule Bills (see Lyons, 1973 and Wicks, 2006). For this reason, it is not included in the analysis in Chapter 5, where I am assessing the impact of the prospects for self-government on legislative behaviour.  53 the current system than preparing for self-government. I expect this when self-government is off the agenda; however if electoral position is still more important when self-government is on the agenda, this would falsify my theory. It is therefore important to look for signs of electorally motivated behaviour when self-government is on the agenda: i.e. attacks on an individual party (the electoral competitor) or credit-claiming, as opposed to anti-system grieving. To reduce the impact of other, unexpected, third variables, I compare the behaviour of the separatist party over two elections or parliaments, as close in time to each other as possible. This means that political culture and institutional structure are held constant, and I can minimize differences in personalities, electoral demographics and the key issues of the day.26  Third, there is a selection effect as all three of the parties studied here won seats in the central legislature and chose to formally take those seats. So, to use Young and Bélanger’s terminology, I am confined to cases at the “democratic” (as opposed to “obstructive”) end of the spectrum – which might limit the degree of anti-system behaviour that we observe. This is an inherent limitation of the research topic, which focuses on how parties engage with institutions that they want to leave. In order to get purchase on this question, we need to actually observe the parties within the central legislature, and in elections where they have to justify their past actions in the legislature. So, we may indeed be explaining and testing separatist behaviour towards the “pro-system” end of the spectrum only. The trade-off is that this allows me to test key predictions from the anti-system party literature. These are Sartori’s argument that these parties will disrupt legislative proceedings and Keren’s hypothesis that they will use the authoritative platform of Parliament to reduce public confidence in the overall political system.                                                   26 In Chapter 4, there was one scenario where the only applicable elections are 18 years apart (the SNP in 1983 and 2001). In section 4.2.1.1, I consider whether differences between these cases might be driving both the independent and dependent variables.  54 2.5 Predictions: how separatists behave over time Campaign strategy when self-government is on the political agenda (tested in Chapter 3): (1) A separatist party is more likely to attack constitutional and institutional arrangements (i.e. the political system) when separation is on the political agenda than when it is off the agenda.  (2) Hypothesis 1 will hold true, regardless of the electoral competitor of the separatist party, or the legislative status of the separatist party.  Campaign strategy when self-government is off the political agenda (tested in Chapter 4): (1) A separatist party is more likely to campaign on the basis of grievances when its           direct electoral competitor is the governing party.   (2) A separatist party is more likely to use a credit-claiming campaign strategy when         its direct electoral competitor is the main opposition party.  (3) Pivotal status will not affect the campaign rhetoric of a separatist party.  Participation in the central legislature (tested in Chapter 5) (1) Separatist Members of Parliament are less likely to participate in any  legislative activities when:        (a) self-government on the agenda, and         (b) the party is not pivotal.   (2) Separatist MPs are more likely to participate in  legislative debates on self-government when: 55 (a) self-government is on the agenda, and  (b) the party is pivotal.  (3)  Separatist MPs are more likely to participate in initiatives to reform the central legislature, and seek benefits for their region when:            (a) self-government is off the agenda and (b) the party is pivotal   These hypotheses are tested in Part II of the dissertation, using data on participation in legislative activities and the framing of issues in parliamentary debates and campaign materials. By comparing data from different points in time, we can assess the impact of the potential for self-government, pivotal status and electoral competition on separatists' behaviour. Furthermore, by comparing the behaviour of separatists to that of other parties in their political system we can test whether the former behave differently and thus merit further study.   56 Chapter 3: Electoral rhetoric and the prospects for self-government Part II of this dissertation tests the theoretical model developed in Chapter 2. It does so by examining the actual behaviour of three separatist parties: The Bloc Québécois (BQ) in Canada, the Scottish National Party (SNP) in the UK and the Irish Parliamentary Party (IPP) in the 19th century UK. This chapter focuses on what I argue is the most important determinant of separatist behaviour, the prospects for self-government, examining the impact it has on electoral campaigns.  As Table 2.2 in Chapter 2 shows, I predict that when self-government is on the political agenda, separatist parties will always take an anti-system stance. We should therefore see campaigns that attack the entire political regime. The separatists will raise problems that cannot be solved by a simple change in government. Rather, the separatist party will use these grievances to argue that its region should govern itself. Similarly, any positive predictions about the future will be tied to the prospect of self-rule.   Self-government is so important to separatist parties that it undermines the impact of the other two independent variables set out in Chapter 2. When self-government is on the agenda, neither legislative influence nor electoral rivalry should make a difference to separatists’ rhetoric. So, even in situations when separatist parties are in a strong position to extract concessions for their region we should still see them attack the system and advocate exit. In these scenarios, therefore, I agree with the anti-system party literature on the likely behaviour of separatist parties.   However, once self-government is off the immediate agenda, the separatist party has to focus on its medium term future within the existing system. This is when my predictions diverge from those of the anti-system party literature. The party’s first order of business becomes electoral survival. It is no 57 longer trying to convert people in its region to the cause of separatism. Instead, it has to widen its base and appeal to groups of voters who are not separatists. Potential supporters fall into two groups. First, there are protest voters who dislike the current government but don’t want to break up the country. Second, there are benefit-seekers who want to see government spending focused on their region. In these elections, we should see the separatist party reducing its anti-system rhetoric and focusing more on what it can offer to voters within the constitutional status quo. This may take the form of attacking the party in government, rather than the wider political system: this kind of rhetoric will appeal to protest voters. Alternatively, the party could target benefit-seeking voters by claiming credit for pressurizing the government to send additional goods and services to the region.   In the first section of this chapter, I explain my case selection and data analysis. I select pairs of election campaigns, in order to compare elections where self-government is on the agenda with those where it is off the agenda. This section also explains how I analyze the parties’ electoral rhetoric: how do I classify it as pro or anti system? Following this are four results sections, each looking at separatist rhetoric in different scenarios. Within each section, I compare an “on agenda” election to an “off agenda” election, to see whether there is a switch from anti-system rhetoric to pro-system rhetoric. The four scenarios depend on the other two variables in my model: whether the separatist party was pivotal in the previous parliament and whether it is competing against the government or opposition party.27  Examining separatist rhetoric in these four scenarios allows me to use pivotal status and the electoral rival as control variables. So, for each pair of elections the only significant difference is the prospect of self-government. I devote the most attention to the first scenario, where the separatist party was pivotal going into the election and was competing against the opposition. As explained in Chapter 2, both of                                                  27 The interaction of these two variables creates four different scenarios: a pivotal separatist party competing against the opposition party, a pivotal separatist party competing against the government party, a non-pivotal party competing against the opposition party and a non-pivotal separatist party competing against the government party. In the final scenario, I actually compare two pairs of cases, so that I can test the theory against both the BQ and the SNP. 58 these conditions encourage a credit-claiming electoral strategy – one that implicitly supports the political system. This scenario is therefore a hard case for the theory to predict. The legislative and electoral conditions mean that we do not expect to see anti-system claims, as the separatist party is well placed to gain concessions from the government of the day. Therefore, if a reasonable prospect of self-government leads to anti-system rhetoric in this type of election, it will have cancelled out the countervailing pressures of legislative and electoral conditions. This would be strong evidence that the prospect of self-government is the most important variable in the model, and that self-government is in fact the primary objective of these parties.  3.1 Case selection and data analysis 3.1.1 Defining “reasonable prospect” of self-government: selecting election campaigns According to the model set out in Chapter 2, the prospect of self-government is the most important driver of separatist parties’ behaviour. This chapter will test whether separatist parties’ electoral campaigns really do change significantly when self-government is off the agenda. The concept of a “reasonable prospect of self-government” refers to the possibility of change to the constitutional status quo, meaning that the region may move further along the spectrum towards independence. There must be authoritative proposals to change the current division of power between the central and sub-national governments (if the latter exist), or for a region to secede altogether. Proposals for a unitary state to introduce a devolved regional government would fit into the category. So would proposals to change a unitary state into a federal one, for a pre-existing sub-national government to gain more powers, or for a region to become an independent country.   Saying these proposals are authoritative means they have been endorsed by the central government of the day or the region’s sub-national government. So, political observers would expect that 59 constitutional change was a reasonable possibility within the next electoral cycle. The separatist party could therefore anticipate that the current electoral campaign may be its last within the existing political system.   The 2x2 matrix in Figure 3.1 sets out the four scenarios in which I am testing separatist rhetoric. Within each scenario, there are a pair of elections from the history of our three separatist parties. Each pair contains an election classified as having self-government on the agenda. This is compared to another election campaign for the same party, where self-government was off the agenda. All of the “on agenda” elections meet at least one of the criteria to be classified as a campaign where there was a reasonable prospect of self-government. The elections in scenarios 1, 2 and 4 are clear examples where self-government was on the political agenda – they are therefore the most straightforward test of whether separatist election rhetoric matches my predictions. In the other two “on agenda” elections (scenario 3 and the BQ election in scenario 4), official commissions had made proposals for self-government. However, neither the central nor regional governments had endorsed these proposals by the time of the election.  By looking at the BQ’s campaign in 1993 and the SNP’s campaign in February 1974,28 we can assess whether the mere existence of proposals for constitutional change is enough to affect separatist party behaviour. Alternatively, we can test whether such proposals only make a difference when they have backing from a government that can put them into practice.  Figure 3.1 also shows comparator elections where there is not a reasonable prospect of self-government. These elections provide a benchmark for judging the separatists’ rhetoric. These are as close in time as possible to the “on agenda” elections. They match the latter on two criteria: whether the separatists had been pivotal in the Parliament prior to the election, and whether the separatists’                                                   28 There were two general elections held in the UK in 1974: in February, when the incoming Labour government established a minority government and in October when it won a majority of 4 seats. 60                                                       Separatist party is pivotal29 Scenario 1 (tested in section 3.2)  Election with self-gvt ON the agenda: IPP, 188630  Comparator election with self- gvt OFF the agenda:                                   IPP, 1895 Scenario 2 (tested in section 3.3)          Election with self-gvt ON the agenda: SNP, Oct 197431  Comparator election with self- gvt OFF the agenda: SNP, 1979   Scenario 3 (tested in section 3.4)  Election with self-gvt ON the agenda: SNP, Feb 1974 (partially)32   Comparator election with self- gvt OFF the agenda: SNP, 1983  Scenario 4 (tested in section 3.5)  Elections with self-gvt ON the agenda: SNP, 201033 BQ, 1993 (partially)34  Comparator elections with self- gvt OFF the agenda: SNP, 2005 BQ, 1997                                                     Separatist party is not pivotal    Figure 3.1 Selecting election campaigns35                                                  29 See Appendix 1 for the distribution of parliamentary seats just before the election, to demonstrate that the separatists were pivotal. This appendix also shows previous election’s results in the separatists’ region, to establish the separatists’ main competitor. 30 The Government of Ireland Bill of 1886 was introduced by the Liberal government and would have established a devolved Irish Parliament. See Aldous, 2006; Kee, 1972 and O’Brien, 1976 for an account of the Bill’s provisions. 31 The Kilbrandon Commission, reporting in 1973, had recommended a devolved Scottish Assembly. These proposals had been endorsed by the Labour government just before the October 1974 election. See Mitchell, 2009 for an account. 32 See footnote 31. In February 1974, the Kilbrandon report had not yet been endorsed by the British government. 33 During 2009, there were two parallel processes considering Scotland’s constitutional future. The Scottish government, controlled by the SNP, issued a White Paper recommending independence. This was followed by a public consultation on independence. See Scottish Government, 2009. The British government established the Calman Commission which recommended greater tax-raising powers and more jurisdictions for the Scottish Parliament. These proposals were accepted by the British Labour government. See Commission on Scottish Devolution, 2009 and UK Government, 2009.  34 The Québec government established the Bélanger-Campeau Commission to consider the province’s future. In 1991, the Commission set out two alternatives: repatriating more powers from the federal government to Québec or holding a referendum on sovereignty for Québec. The Québec Liberal government set up parliamentary committees to report on both tracks, without endorsing either. See Cardinal, 2005 and Young, 1999 for an overview of the Commission’s work 35 Sources: Author’s analysis of data found in: Craig 1984; Craig 1989; Dod’s Parliamentary Communications, 2005; Henig and Baston, 2002; Parliament of Canada, 2014; Rallings and Thrasher, 2000; Electoral Commission (UK), 2006; UK Parliament 2014; Walker, 1978. Electoral competitor: opposition party Electoral competitor: government party 61 primary electoral competitor was the government party or an opposition party.36  3.1.2 Measuring election rhetoric: the dependent variable        To measure whether separatists’ electoral campaigns are pro-system or anti-system, I hand-coded the leader’s forewords in manifestos, candidate speeches and campaign leaflets from the selected election campaigns.37  Election manifestos and leaflets can be thought of as a contract between a party and voters. The party leadership commits itself publicly to a policy programme over the next legislative period and reports back on what it has achieved over the previous period. Manifestos provide a useful measure of overall campaign strategy because they allow parties to explain and justify policy proposals in detail and to draw linkages between their positions on a variety of issues. They are free of the institutional constraints of the legislature (where the government party usually controls the agenda), they do not have the cost or space limits of advertisements or social media and they are not open to the misinterpretation that may occur in media interviews. A manifesto can therefore be thought of as the party elite’s definitive statement of its position at a particular point in time.38    I coded each sentence of the manifesto, speech or leaflet separately. Each sentence was placed in one of four categories, as set out in the 2 x 2 matrix in Figure 3.2.39  The y axis in Figure 3.2 shows whether the sentence is claiming that the state of affairs in the separatists’ region is positive or negative. The x axis measures whether the claim is pro- or anti-system. If the positive claim is made in order to praise                                                   36 These factors are independent variables in their own right in my theoretical model: see Chapter 2. Their impact on electoral rhetoric will be tested in Chapter 4. 37 Appendix 2 sets out the data sources used to find manifestos for these campaigns. It also explains the variance in data availability for different time periods, the use of leaflets and verbatim reports of speeches as substitutes for manifestos and how I dealt with problems of translation and missing data. 38 See Marland et al, 2012 for an argument that manifestos tell us about the party hierarchy’s priorities: “Election campaigns involve the strategic allocation of finite resources in a competitive marketplace.” (p. 73). 39 This section give a brief description of what these four categories mean, and how they map on to the key theoretical concepts of anti-system claims and pro-system claims. Appendix 2 sets out the detailed steps involved in coding each sentence. 62                                                                 Positive claims Credit-Claiming:  Positive claims which are attributed to the separatist party’s actions within the current system, the current government or the overall political system. OR  Reasons to vote for separatists, premised on the benefits they have brought to the region. OR  Positive claims about the future, premised on a substantive policy that will be implemented within the current system.             The Case for Self-Government  Positive claims about the future which are attributed to increased self-government  OR   Reasons to vote for separatists premised on bringing about constitutional change  OR   Positive claims about the future premised on a policy of increased self-government                           Anti-Government Grieving  Negative claims which are attributed to the party in government, or specific politicians or policies.                        OR  Reasons to vote for separatists premised on defeating a specific party or candidate. OR  Negative claims which are linked to a substantive policy that will be implemented within the current system.            Anti-System Grieving  Negative claims which are attributed to the overall political system or country. OR  Reasons to vote for separatists premised on bringing about constitutional change OR  Negative claims which are linked to a policy of increased self-government                                                                 Negative claims Figure 3.2 Categorizing campaign messages  Pro-system claims Anti-system claims 63 the current political system, or to urge voters to support stability, or to suggest the continuation of a substantive policy, then it is classed as pro-system. However, a positive claim could also be serving an anti-system purpose, if it is linked to the benefits of greater self-government or separation.   Similarly, negative claims may be either pro or anti system. They implicitly support the overall political system if they attribute problems to the government of the day, or to another party or to a policy. These types of claims are the same as we would see from mainstream opposition parties. They are about a clash of ideology or a debate over which party would make a more competent government, but do not threaten the overall political regime. Conversely, negative claims which are used to build a case for self-government or outright separation are clearly anti-system, and would only be made by separatist parties. Categorizing each sentence in each manifesto in this way allows us to measure the relative weight given to pro and anti-system messages in any given election campaign. Crucially, we can also compare different elections over time. In this chapter, I use this temporal comparison to test whether a party’s rhetoric becomes significantly more anti-system when self-government is on the agenda.   3.2 Results for scenario 1: a pivotal separatist party competing against the opposition I start with the cases which pose the most difficult test for my argument: the Irish Parliamentary Party (IPP) in the 1886 and 1895 elections. In both campaigns, the IPP had held the balance of power in the previous parliament. Furthermore, the governing Liberal party did not run for seats in Ireland, so the IPP was almost exclusively competing against the Official Opposition party. This was the Conservative party, which from 1886 onwards had allied with former members of the Liberal party who wanted to keep the United Kingdom as a unitary state. These “Liberal Unionists” formed a legislative block with the Conservatives, opposing Irish Home Rule. Candidates from this alliance usually ran under a 64 “Unionist” party label in Irish seats (Walker, 1978). According to the model laid out in Chapter 2, both these factors push a separatist party towards a credit-claiming electoral strategy, which implicitly supports the political system. Pivotal status gives separatist parties opportunities to pressure the government of the day into concessions for their region. This allows a credible claim that the separatists’ presence in the central parliament is good for the region. The second factor (separatists’ competition with the Official Opposition) alters the incentives of the government party. If the separatist party wins seats, it does so at the expense of the opposition. So, in this scenario the government party has both electoral motives and legislative pressures to send goods and services to the separatist region.   If, however, the prospect of self-government overrides these consideration, then we should see a stark difference between the IPP’s electoral behaviour in 1886 and 1895. In the first of these elections, Gladstone was Prime Minister and had introduced a Home Rule Bill in April 1886. This was endorsed by the Irish Nationalist movement both in and outside of Parliament, with an editorial in the Nationalist weekly newspaper United Ireland calling it “the greatest event of the century for Ireland” (10 April, 1886). The Bill would have established a two-house legislature for Ireland, though with certain jurisdictions and overall fiscal responsibility reserved to Westminster (O’Brien, 1976, 31-32). It therefore gave the Irish some powers of self-government but stopped well short of independence.   Nevertheless, the Bill was opposed by those who rejected any challenge to the Westminster Parliament’s sovereignty, and was defeated by the Unionist alliance in June of that year (Mitchell 2009, 7-12). This set the stage for the election of the following month, with both the governing Liberals and the IPP campaigning on the basis of a re-introduction of the Home Rule Bill in the next Parliament. Contemporary news reports show that political observers in Ireland viewed Home Rule as still being clearly on the agenda. After the defeat of the Bill, one newspaper told its readers: 65 “The progress of the cause of Irish Nationalism has suffered a check but it has at the same time reached a point at which five years ago it would have been deemed simply impossible to reach within that period of time. The Irish question now is to the forefront of politics and the defeat of last night cannot relegate it to a minor place.” Cork Examiner, 8 June 1886, cited in Kee, 1972.   By the time of the 1895 election, the government had indeed introduced a second Home Rule Bill. However, the Bill had been defeated in the House of Lords in 1893. This second defeat was seen as being far more final by the politicians of the day. The House of Lords was viewed as an insurmountable constitutional obstacle (Wicks, 2006) and Gladstone had retired after the Bill’s defeat. The new Liberal leader, Lord Rosebury, was not a Home Rule supporter, and gave it no more parliamentary time (see Kee, 1972, 422 for Rosebury’s views). Furthermore, the Irish Nationalist MPs were themselves split into two factions.40   So, if the theoretical model is correct, we should see a stark difference between Irish Nationalist rhetoric in these two elections, with the 1886 election being far more anti-system in tone. Once Home Rule is off the immediate agenda in 1895, other factors should affect the IPP’s behaviour. The Nationalists’ pivotal status and their electoral struggle with the opposition party should encourage them to make pro-system claims to justify their presence at Westminster. This section examines that hypothesis by first examining whether the candidates themselves were aware of these two strategic factors: did they understand the leverage that pivotal status gave them and how did they characterize the opposition party when talking to their voters? Second, I look at the big picture, comparing data on candidate speeches to see if they were significantly more anti-system in 1886 than in 1895. Third, I look at the detail of the speeches to show how IPP candidates say they used their pivotal power. When                                                  40 See Kee, 1972, 410-413 and O’Brien, 1976, 88 for the fall of the IPP’s leader, Charles Parnell in 1890, and the subsequent split between the Parnellites and anti-Parnellites. 66 seeking concessions did they prioritize self-government, or did they focus on substantive benefits for Ireland that could be obtained without Home Rule?  3.2.1 Do candidates explain their strategic position to voters? In both elections, IPP candidates made it clear that they viewed the Unionists as their primary opponent. They also linked the Unionists to the Conservative party, which had been the Official Opposition in the previous parliament. For example, in 1886, the IPP candidate for Monaghan North contrasted the two British party leaders as follows:  “ We have, as I have said before, 200 followers of Mr. Gladstone pledged to support our cause, and on the other hand we have Lord Salisbury proposing to give us twenty years of coercion . . .” Patrick O’Brien, Monaghan North, July 14 1886, sentence 13.  So, O’Brien paints the Conservative leader, Lord Salisbury, as a threat not only to the IPP, but to all Irish people. “Coercion” was a catch-all term used by Irish Nationalists to refer to any British government legislation which suspended civil rights, particularly when it was aimed at Ireland alone rather than the whole of Britain (Kee, 1972, 377-80). On the other hand, Gladstone, the Liberal Prime Minister, is shown as being a supporter of “our” (the Irish Nationalists’) goals. He is also described by O’Brien as being able to command enough support to actually deliver on his promises.    The same positioning can be seen in 1895, in this report of a speech by the Nationalist candidate for Country Down, South: “. . . on the day of the polling every man would come forward, and that South Down would give such a fright to the Tories that they would never again put forward a candidate to oppose the National will.” Freeman’s Journal report of speech by Michael McCartan, Down South, July 11 1895, sentence 30  Here, there is not the same closeness to the Liberal party as we see in 1886 – as noted above, the new Liberal leader, Lord Rosebury was wary of the Irish Nationalists. However, the antipathy to the 67 Conservatives remains: they are depicted not only as simple electoral rivals, but as a party that McCartan would like to drive out of the constituency altogether.   For their part, Unionist candidates reinforced this understanding of the electoral competition. In both campaigns, their electoral rhetoric was hostile to the IPP. Furthermore, Unionists made it clear that they would support a Conservative (rather than Liberal) administration in the House of Commons. For example, the Unionist candidate in County Antrim East links his Unionist candidature directly to Lord Salisbury’s Conservatives: “In the almost certain belief that the Unionist . . . party will be successful at the polls, it will be my privilege as your Representative, to give unqualified support to Lord Salisbury’s Ministry . .” Election Address of James McCalmont, Antrim East, Belfast News-Letter, July 11 1895  Unionist candidates in Armagh Mid, Antrim North, Belfast North, Belfast South and Belfast South made similar statements about providing parliamentary support to the Conservatives.41    In addition to making their electoral position clear, Nationalist candidates also acknowledged the legislative power they had enjoyed in the previous parliament. During the 1886 campaign, the candidate in Monaghan South noted that the split within the Liberal party had: “ . . . left the balance of parties on every issue but one absolutely at the disposal of  the party led by Mr. Parnell.” Freeman’s Journal report of speech by Joseph McKenna, Monaghan South, July 10 1886, sentence 30  The one issue not controlled by Parnell was Home Rule itself, but McKenna here clearly recognizes the potential for gaining concessions in other issue areas. We see similar sentiments in 1895 when Thomas Healy regrets that the IPP did not take full advantage of its pivotal power before the election. In doing                                                  41 See the Belfast News-Letter, 9 July 1895, 11 July 1895, and 13 July 1895. 68 so, he acknowledges the leverage the Nationalists had, even if he is not impressed with how his party used it: “. . . he sincerely hoped that the Irish Party would learn by lessons of the last three years and would put more pressure on whatever party was in power. It was all very well talking about having a Government in the hollow of your hand, but you wanted to squeeze them to get anything out of them . .” Freeman’s Journal report of speech by Thomas Healy, Wexford North, July 8 1895, sentence 9-10  Furthermore, in both elections, candidates mused on the paradox of their position as Irish Nationalist candidates who were asking voters to send them to the Westminster Parliament. In 1886, one candidate was optimistic that this would only be a temporary state of affairs: “He hoped that in two or three years’ time he would again be seeking their suffrages, not to send him to the Imperial Parliament, but to the old house at home [The Irish Parliament in Dublin which was abolished by the 1801 Act of Union]” Freeman’s Journal report of speech by Matthew Kenny, Tyrone Mid, Jul 10 1886, sentence 11  By 1895, the tone has changed, with MPs not making promises of short-term constitutional change. Instead they point to personal sacrifices in going to Westminster, using this as proof of their commitment to their constituents: “Well it was their [the people at the meeting] duty to go to business, and that business was once more to send to Parliament men who would truly represent the people. . . If they [the speaker’s audience] only knew the truth they would believe him when he said that it was not a pleasant thing to serve long in Westminster; but when there of course a man should do his utmost, and he had not spared himself.” Freeman’s Journal report of speech by Michael McCartan, Down S, July 11 1895, sentences 18 and 20.  These statements might be seen as mere rhetoric, designed to impress voters with the diligence of the Nationalist MPs. However, the Unionist MPs from Irish constituencies had the same travelling distances and cultural differences to contend with at Westminster. Yet, their speeches show neither the 1886 Nationalists’ view that being a Westminster MP was a short-term position, nor the 1895 Nationalists’ view that it was an unpleasant necessity. In 1886, for example, one candidate stated his enthusiasm for being able to vote on Irish affairs in Westminster: 69 “It was a great satisfaction to me that I was able to assist at the rejection of Mr. Gladstone’s proposals in the division which has brought about the Dissolution and I trust to have the pleasure of uniting with the patriotic Members of the new Parliament in sealing the fate of any similar measure the Prime Minister might propose.” William Ewart, Belfast North, Belfast News-Letter, June 29 1886.  Other Unionist candidates praised the institution of Westminster itself. They argued that it was an impartial tribunal for resolving Ireland’s religious conflict (Matthew Megaw, Tyrone East, 1886), that the Irish were “fully and fairly represented” there (Henry Mulholland, Londonderry East, 1886) that it was “willing and eager” to make reparations for injustice (Ernest Hamilton, Tyrone North, 1886) and that it was an ancient institution which should not be attacked (James McCalmont, Antrim East, 1895).42  It was also common for Unionists to sign off their election addresses by stating that it would be an honour to represent their constituencies in London.43  This kind of claim is of course familiar to most electoral campaigns but it is in clear contrast to Nationalists’ references to Westminster. This underlines the separatist ambitions of the latter. However it also highlights the changes we see between the Nationalists explanation of their role as MPs in 1886 (when they expected to leave Westminster shortly), and in 1895 when this was a distant prospect.   So, we know that at least some Nationalist candidates appreciated their strategic position going into both these elections, and were happy to share this with their voters. First, they made it clear that they were no friends of the Opposition party (the Conservatives) and in some cases, candidates went out of their way to praise the incumbent government and particularly the Prime Minister, Gladstone. The Nationalist candidates also spelled out the power they enjoyed in the previous parliament, which could have been used to gain concessions. Furthermore they reiterated that (in contrast to the Unionists) they                                                  42 See the election addresses for these candidates, in the following editions of the Belfast News-Letter: June 25 1886, June 26 1886, June 29 1886 and July 11 1895. 43 See, for example, the election addresses of James McCalmont, East Antrim, Belfast News-Letter, July 11 1895 and William Johnston, South Belfast, Belfast News-Letter, July 9 1895. 70 do not feel comfortable in a non-Irish legislature. In this way, they drove home the message that they were only there to represent their voters’ best interests, rather than gain personal perks or access to office. All of these factors make it more likely the Nationalists would use credit-claiming rhetoric. In arguing that their presence at Westminster allowed them to win concessions for Ireland, they would be implicitly supporting the status quo. These elections are therefore a difficult test for my theoretical model. We would expect to see a pro-system campaign in 1895, when Home Rule is off the agenda. That leaves 1886 – will the prospect of self-government be enough for the IPP to ignore all the incentives to credit-claim?  3.2.2 Campaign messages: pro- or anti-system? Table 3.1, below, compares all the candidates speeches made by IPP candidates in the two elections, looking at the balance between pro-system and anti-system claims. An example of an anti-system claim is the following quote from an IPP candidate in the 1886 campaign. This sets out a negative description of conditions in Ireland, and crucially, attributes these to the overall political system – the “English” government:   “The English government produced by their 700 years of misrule a poor, discontented, and unhappy people – an Ireland deprived of all the advantages that go to make a country prosperous and enlightened, an Ireland robbed of all the sources of industry, and the receiver poorer than he was before.” Freeman’s Journal transcript of speech by Richard McNabb, Down North, July 13 1886, sentence 2   The quote over the page from an IPP candidate in the 1895 election also suggests that there are serious problems in Ireland, as he is advocating for land reform. Here, however, the candidate does not blame the political system. Instead, he sets out in detail the reforms he hopes that Irish votes will secure at Westminster – so he is stating how he would use the current, British political system to address Ireland’s problems:   71 Table 3.1 Categorizing campaign messages, IPP 1886 versus 1895  1886 Self-government ON agenda 1895 Self-government OFF agenda Pro-system sentences (Credit-claiming and Anti-government grieving)  92 sentences (29.8% of campaign) 79 sentences (74.5% of campaign)  Anti-system sentences (Positive case for Separation and  Anti-system grieving) 217 sentences (70.2% of campaign) 27 sentences (25.5% of campaign) Total for campaign 309 sentences  106 sentences  N=415 sentences Missing cases (sentences coded as neutral): 79   “What, in my opinion, would be justice to you and fair to the landlord would be to have a sweeping land measure passed – a Land Bill that would reduce the rents, taking into account the amount of their improvements, and having reduced the rents very considerably, very much from what they are now, we should try and pass a Land Purchase Bill at no higher than ten years’ purchase. I shall not be absent from a division that your interests or the interests of your country is at stake.” Freeman’s Journal transcript of speech by Patrick Doogan, Tyrone East, July 5 1895, sentences 7-8   Table 3.1 looks for these two types of rhetoric across speeches from every IPP candidate. It shows a clear pattern that fits with the model’s predictions. In 1886, when the Prime Minister of the day supported Home Rule, 70.2% of the speeches comprised anti-system messages. In 1895, we see a complete turnaround, with three-quarters of the campaign rhetoric now being devoted to pro-system claims. The difference in the distribution of messages between the two campaigns is significant at the 99% level.44 We can also calculate the strength of the association between, on the one hand, the 1886 campaign and anti-system messages, and on the other, the 1895 campaign and pro-system messages. Using a phi-adjusted statistic, we find that the level of association is 0.567, which is considered to be a moderate relationship (Bohrnstedt and Knoke, 1982). Put another way, the odds of a sentence from an                                                  44 The chi-square statistic 65.2539, with a p-value of 0.000. 72 1895 speech supporting the current constitutional system is 6.9 times that of a sentence from an 1886 speech.45    This is quite an encouraging result, given the number of other factors (besides the prospect of self-government) that can affect election campaigns. It is also worth noting that these elections pre-date the era of centralized campaigns and strict discipline over every statement made by a party’s representatives. So, there are likely to be more idiosyncratic messages in the IPP’s campaign material than for 20th and 21st century parties. Localized issues and personal animosities are more likely to feature in individual speeches not vetted by a party administration than they are in a manifesto written by those at the top of a party hierarchy. Despite this, we still see a clear difference between the campaigns, suggesting that a range of candidates decided to shape their messages differently in 1895: when they knew they would probably be part of the British political system for the foreseeable future.46    3.2.3 How does the IPP use its pivotal status? As shown in section 3.2.1, Nationalist candidates explained their parliamentary leverage to their constituents. We can therefore imagine constituents wanting to know how their MP (or the party) had used this power. Examining how candidates addressed this point is useful for two reasons. First, it gives us information on what the candidates’ priorities were. This allows us to test the model’s arguments about a separatist party’s preferences: that self-government will be the primary objective, but when it is off the political agenda, secondary preferences come into play. These may be for electoral survival, other policy objectives or goods and services for the region. Second, and most importantly, candidates’ statements show us how they want voters to perceive them. The opportunities                                                  45 Conversely, the odds of a sentence from 1886 conveying a pro-system message are 0.14 times the odds of an 1895 sentence. 46 Appendix 3, Tables C.1 and C.2 give further detail on the speeches used in each of the IPP campaigns. 73 provided by pivotal status are an electoral tool, and these speeches show us how the party deployed that tool.   In 1886, several candidates make it clear that pivotal status has opened up the chance for a Home Rule Bill. For example,  “An opportunity in the present grave and critical moment in Irish affairs now offers [sic] for changing this [the Act of Union, 1801], and for returning to Parliament men who by aiding the efforts of the two great leaders of English and Irish thought [Gladstone and Parnell], will bring back to our shores the Legislature of which this Nation was so foully despoiled.” Hugh Herbert Johnson, Dublin University, 1886 election address, sentence 4.  Furthermore, candidates explicitly stated that Home Rule outweighed any other consideration, including further land reform in Ireland, which had been the rallying call for a range of Irish movements since the middle of the century.47 Notably, this came from candidates in rural areas such as Country Tyrone, where farming was the primary means of survival. These are the constituencies where we would expect land reform to be particularly salient to voters. So, the fact that candidates in these areas are prioritizing Home Rule underlines the importance of self-government to the party: “In reference to the land question Mr Kenny said that all other reforms were of minor importance in comparison with the settlement of Home Rule, and a Land Bill that would confer any benefit on the farmers would not be passed until one was passed by the Parliament in College green [the site of the former Irish Parliament, abolished in 1801].” Freeman’s Journal report of speech by Matthew Kenny, Tyrone Mid, July 10 1886, sentence 3.  In other words, to a rural candidate like Kenny, land reform is the reason for concentrating on Home Rule. Tellingly, this argument is also made by William O’Brien. O’Brien was one of the leaders of the Irish Land League, who only entered politics in order to secure three things for tenant farmers: fixity of tenure, lower rents and the ability to sell their interest in the land, receiving compensation for any                                                  47 See Kee, 1972, Part 4, Chapter 2 for a history of land tenure and reform initiatives in Ireland. 74 improvements they had made. He was not a natural politician, and indeed decided not to run for Parliament in the 1890s, by which time rents had stabilized and there were fewer tenant evictions.48  However, in 1886, even such a committed supporter of land reform urges the prioritization of Home Rule: “It is not Home Rule that has caused the misfortunes of the tenant-farmers of Ireland – it is Home Rule that will cure them.” Freeman’s Journal transcript of speech by William O'Brien, Tyrone South, June 26 1886, sentence 9.  Finally, one candidate sums up the primacy of self-government by basing his campaign solely on exit from the Westminster system: “ I appealed for your suffrages six months ago on the single plank of the restoration of Ireland to her native Parliament. You by your votes indicated your devotion to that principle. We perhaps could scarcely hope that within so short a time we should find ourselves on the eve of obtaining that Parliament. I myself believed that probably the struggle would be a longer one and more difficult one than now appears likely.” Freeman’s Journal transcript of speech by Edmund Dwyer Gray, Dublin St Stephen's Green, June 29 1886, sentences 3-6.  It is revealing to compare these statements to those of the Unionist candidates in the same election. The latter agree that Home Rule is the primary issue, and ask for electoral support to defeat the measure.49  However, they focus on substantive issues too, pointing to the positive things that Westminster could do for Ireland. The Unionist candidate for West Belfast points to his work in the previous parliament to secure a local bankruptcy court for Belfast, and is optimistic that this will be possible if a Conservative government returns to power (Belfast News-Letter, July 3 1886). Another candidate has more extensive ambitions:                                                    48 See Joseph V. O’Brien’s biography of William O’Brien, 1976 49 See, for example, the election addresses of William Ewart, Belfast North (Belfast News-Letter, June 29 1886) and Henry Mulholland, Londonderry East (Belfast News-Letter, June 25 1886). 75 “By opening up the country by an extended system of Light Railways, by stimulating the Sea Fisheries, and by the establishment of Government industries and manufactures throughout the country . . . This, however, can only be achieved by the help of English funds and English enterprise, and it will be my object to do all in my power to induce the English Government to take such steps as may be the means of speedily restoring prosperity and plenty to Ireland.” Election address of Ernest Hamilton, Tyrone North, Belfast News-Letter, June 29 1886. Emphasis added.  This is a clear statement of the economic case for maintaining the Union between Ireland and Britain. Unlike the Nationalists, the Unionist candidates welcome “English” intervention in Irish affairs, arguing this is the only way Ireland can develop. Here, the Unionists are focusing on positive things that can be achieved in the future, rather that stating that Ireland is in a good position now. This makes sense, given that the Unionists’ rivals (the Liberals) were in power in the previous Parliament. So, in line with my theory, the Unionists do not want to praise the government of the day. However, they rebut the Irish Nationalist argument by claiming that, under a different government, the British political system can bring great benefits to Ireland. Furthermore, the Unionist candidate quoted here sees it as his job to pressure the next government to do just this.  Turning to 1895, when Home Rule is no longer on the agenda, we see a shift in the rhetoric of the Nationalists (but not the Unionists). Non-constitutional issues take priority, particularly when candidates are explaining how they will use their position in Parliament. For several Nationalists, land reform is again centre stage, as seen in this quote: “No farmer in this room, in this province of Ulster, or in Ireland, will have a warmer wish for their interest than I have, and no-one will be more anxious to procure for them a good Land Bill. No labourer in Ulster or in Ireland will want a friend while I am in Parliament.” Freeman’s Journal transcript of speech by George Murnaghan, Tyrone Mid, July 8 1895, sentences 9-10  This is remarkably similar to the claim by a Unionist candidate in the same election, though, unsurprisingly, he states that it is a Unionist-Conservative government that will deliver land reform: 76 “The Unionist government [which he anticipates being elected in the 1895 campaign] contains men who represent both the Landlords and the Tenants, and I have every confidence in the willingness and capacity of such a Government to improve the Land Purchase Acts and to make all necessary Amendments in the Land Acts in a way which will give general satisfaction, without injustice to any interest.” Election address of Dunbar Plunket Barton, Armagh Mid, Belfast News-Letter, July 13 1895.   The above two quotes are examples of promises of future action, focused on what the candidate will do if elected to Parliament. Separatist participation in the British political system is even more apparent when candidates use credit-claiming: explaining what they have already achieved at Westminster. Michael McCartan, the veteran Nationalist MP for southern County Down devotes several paragraphs to his committee work on various Land Reform Bills (Freeman’s Journal, July 11 1895; see in particular sentences 10-13). This speech set out in detail how the machinery of the House of Commons could be used to study a subject of great concern to his constituents, and how he himself had participated in these institutions and co-operated with the government of the day.   This is in contrast to the Nationalists’ 1886 statements about land reform, which argued that this was not something that could be achieved at Westminster, thus necessitating Home Rule. Interestingly, McCartan ends his 1895 account by explaining that the Land Bill failed because the Liberal government was defeated on a motion of no confidence on an unrelated topic. He specifies that the motion was brought by TW Russell, a Conservative Unionist MP running in South Tyrone (Walker, 1978). So, instead of blaming the overall political system for the failure of land reform, McCartan is blaming a specific individual – one of his party’s electoral rivals. This is the equivalent of anti-government grieving, albeit aimed at the opposition party. The Nationalist candidate is therefore behaving like any other mainstream politician, attributing failures and problems to other parties, not institutions. In particular, in 1895, the Nationalists are making the same rhetorical claims as the pro-system party in their own region. This is what my model predicts, as this is an election campaign where 77 the Nationalists know they will be interacting with the same (British) electoral rivals for the foreseeable future.   In 1895, we also see credit-claiming on subjects other than land reform. This suggests that the Nationalists prioritized a range of substantive policy areas once Home Rule was not an immediate prospect. For example, one candidate pays tribute to his Nationalist predecessor in the constituency by referring to investment secured for the local area. The claimed achievements of the previous MP sound very similar to the benefits promised by Unionist candidates in 1886 (see the quote from Ernest Hamilton, above), focusing on transport infrastructure: “He regretted that Mr McGilligan had thought it wise to retire from a field in which he was an assiduous worker, and he had rendered valuable assistance in connection with various undertakings in the county – such as the light railways and other matters – in which he did the best for the interests of the county and of his constituency.” Freeman’s Journal report of speech by Jeremiah Jordan, Fermanagh, July 9 1895, sentence 6  Candidates also used their speeches to set out their record as good parliamentarians. They therefore made the general argument that their presence, and active participation, at Westminster was beneficial for their constituents: “I say that as long as I have health I shall sit in my seat in the House of Commons and vote in every division I possibly can. . . I say here today that my unceasing efforts in Parliament will be to promote the good of my country.” Freeman’s Journal transcript of George Murnaghan, Tyrone Mid, July 8 1895, sentences 3 and 6 and “He had never missed a vote in the House of Commons on any matter affecting Ireland.” Freeman’s Journal report of speech by Eugene Crean, Ossory, July 22 1895, sentence 11.  The 1895 Nationalist quotes considered so far are all from anti-Parnellites. However, among the minority Parnellite faction, we see similar sentiments. The Parnellite candidate for Dublin South quotes 78 Thomas Healy (an anti-Parnellite) in expressing disappointment with what pivotal status bought the Irish Nationalists: “After three years of close alliance with the Liberal Party – after three years of British reforms enacted by grace and favour of seventy-one Irish votes – seventy-one Irish MPs come back to you (to borrow Mr Healy’s own expression at a recent meeting) “empty handed” . . . Edmund Haviland-Burke, Parnellite, Dublin South, election address 1895, sentence 11  However, in listing the policy areas where he wanted to see concessions but got none, Home Rule is merely one of many. It is not listed first, nor is it given any rhetorical prominence: “ . . .without Amnesty, without reinstatement of evicted Tenants, without justice to the Christian Brothers, without Home Rule, without any one of the dozen National benefits so confidently promised to you as the result of loyalty to British Liberalism.” Edmund Haviland-Burke, Parnellite, Dublin South, election address 1895, continuation of sentence 11  Furthermore, his solution to this lack of progress is that the Irish Nationalists should continue to participate in votes in the House of Commons, and continue to co-operate with British parties, though being less choosy about whom they work with: “The policy on which I appeal to you is the policy of making “Ireland Block the Way”, of supporting any British Party that will serve our purposes: of opposing any British Party that will make the Irish Question wait on any British reform.” Edmund Haviland-Burke, Parnellite, Dublin South, election address 1895, sentence 12  So, in both of the Irish Nationalist factions in 1895, we see a sophisticated appreciation of the party’s leverage. In this period when politicians know Home Rule will not be achieved soon, they tell their constituents that they will continue to work at Westminster, rather than boycott it or set up alternative institutions in their own region. This is similar to the promises made by Unionists, but in contrast to the Nationalists’ own campaign in 1886. In that election, the IPP was on friendlier terms with the Prime Minister of the day, yet all their discussion of how to use pivotal status was focused on anti-system claims. Home Rule was the key objective, drowning out other concerns. However, once Home Rule 79 was off the agenda in 1895, the Nationalist candidates returned to other policy concerns such as land reform and were keen to show what they could achieve for their constituents within the British political system.    This section took an in-depth look at how the theoretical model worked in “hard” cases, where the separatists had the legislative opportunities to claim concessions and electoral reasons to claim credit for these concessions. The next three sections extend that work by testing the model in the three other legislative and electoral scenarios set out in Figure 3.1. These include two scenarios where the separatist party is competing with the government of the day, so credit-claiming is unlikely to be a winning strategy. This is because benefits-seeking voters will simply flock to government candidates, not the separatist party. In scenario 3, the separatists are competing against an opposition party (as the IPP was) but are not pivotal, making claims they gained concessions somewhat less credible. So, in all these cases, credit-claiming is a less attractive strategy than it is in scenario 1. However, messages which implicitly support the political system are still feasible: a separatist party can deploy anti-government grieving (particularly when competing against the government). In doing so it behaves in the same way as opposition parties, as it is arguing that the solution to its region’s problems is to simply change government, or policy, not to leave the country altogether. So, given that there is still scope for pro-system messages, I need to test whether the prospect of separation make a difference in these cases. In these elections, do we still see a shift from anti-system to pro-system rhetoric once self-government is off the agenda?   3.3 Results from scenario 2: a pivotal separatist party competing against the government To apply the model to a scenario 2 case, I use the Scottish National Party’s election campaigns in October 1974 and 1979. Prior to both elections, the incumbent Labour government had a minority of 80 the seats in the House of Commons. In Scotland itself, Labour was clearly the dominant party (see Appendix 1, Table 1), meaning that if the SNP was looking to pick up more seats, or defend the ones it had, it was usually Labour that it had to challenge (see also McLean, 2004 on the Labour-SNP rivalry).   There were two background issues in the 1970s which directly affected Scotland. First, North Sea oil was discovered in the early 1970s. This formed the basis for much of the SNP’s case for outright independence, as seen in the slogan “It’s Scotland’s Oil” (Mitchell, 2009, 118). The party used the existence of oil to argue, first, that an independent Scotland was economically viable, and second, that remaining in the UK meant that Scotland’s resources would be siphoned off to fund services in the rest of Britain, chiefly England. Second, the UK had entered the European Economic Community (EEC) in 1972, under Edward Heath’s Conservative government. This had been confirmed by a statewide referendum held by Harold Wilson’s Labour government in 1975. The EEC was therefore a salient issue at both elections in 197450 and at the 1979 election. In Scotland, it had a particular impact on the fishing industry, as British waters were opened up to fleets from other EEC members, and British fisheries also had to comply with European quotas for catch sizes (Bennie 1997, 13).  The key difference between these elections is the issue of devolution. The government-endorsed Kilbrandon Commission reported in 1973, recommending a separate Scottish Assembly, with some spending powers (Mitchell, 2009). The Labour party was initially skeptical but after it won the February 1974 election, there was a change of policy. Wilson’s government issued a White Paper on implementation of devolution in September 1974, the day before dissolving Parliament (Mitchell,                                                  50 Edward Heath called an election in February 1974, which resulted in a hung parliament and a minority Labour government. As the incumbents were Conservatives (i.e. not the SNP’s primary rivals), and as Heath had a parliamentary majority, the February 1974 election is classified as a “scenario 3” election and is dealt with in section 3.4. The incoming Labour Prime Minister, Harold Wilson, held another election in October of that year in a successful attempt to gain a parliamentary majority. However, by the time of the 1979 election, this majority had been whittled away through by-election defeats and defections, so that the government was once again in the minority.  81 2009, 120). Self-government was thus clearly on the agenda in the October 1974 election. However, in 1979, a referendum on Scottish (and Welsh) devolution failed to reach the threshold of required votes. The election followed shortly after, so it was clear that for the next electoral cycle, at the very least, Scotland would continue to be governed by the Westminster Parliament.  In the October 1974 election, we see the SNP using the “Scotland’s Oil” slogan to raise grievances, in which Scotland is cast as the victim of London. Geographic terms are used here, as opposed to named political parties, politicians or the government of the day, so this is anti-system grieving: “Without self-government this money will go straight to London. We’ll be lucky if we get a fraction of it. After paying for a third London airport, Concorde, an extended Polaris programme, a new London motorway system, the Common Market levy and all the rest, there will be very little left for us.” SNP, October 1974, Leaflet 3, Sentences 6-8 Furthermore, the quote begins by presenting self-government as the solution to this problem. Earlier in the same leaflet, we see oil being used to actually make the case for separation, spelling out what Scotland could gain if it did not have to share the oil with anyone else:  “At the very least we’d get £825 million more every year. That’s half as much again as the total of all public expenditure in Scotland today.” SNP, October 1974, Leaflet 3, Sentences 2-3.  Oil revenues also feature in the 1979 campaign, but here the SNP attacks current levels of oil production. Its proposed solution does not mention any form of self-government nor does it pit Scotland against the rest of the UK. Instead, the party argues for a substantive policy (conservation) which could be implemented without constitutional reform: “While other oil producing countries are conserving supplies so that they will last for 100 years or longer, our oil will be gone in 35 years. Our children are going to be poor because we are being greedy.” SNP, 1979, Leaflet 17, sentences 43-44  82 So, on this issue, we see a shift from anti-system grievances and the economic case for independence (in 1974) to a call for the existing government to change its policy (in 1979). There is a similar pattern in the SNP’s arguments about the EEC. When devolution was on the agenda, the party makes the following argument against the UK’s membership: “Although a majority of Scottish MPs voted against it, the Westminster government forced us into the Common Market.” SNP, October 1974, Leaflet 7, sentence 1.   This fits with Chapter 2’s prediction that separatist parties will deploy procedural arguments, as they are essentially concerned with who makes decisions. Here, the party’s rhetoric pits Scotland against Westminster, in the same way as its 1974 statements about oil. The objection to EEC membership is that Scotland’s elected representatives were against it. The party does not see a Westminster majority made up of non-Scottish MPs as having any legitimacy to bind Scotland.   In the same campaign, there were leaflets which made the substantive case against the EEC, spelling out negative effects for Scotland’s rural areas: “With Banffshire’s beef farmers facing heavy losses and dairy and pig farmers going out of business, the position has become desperate.” SNP, October 1974, Leaflet 5, sentence 4 However, once the leaflet moves on to providing a solution, the farmers’ difficulties are again linked to self-government. First, the loss of farms is framed as the inevitable consequence of EEC membership. Second, self-government is claimed to be the only way in which Scotland can control whether it remains part of the EEC, and if so, on what terms:51   “So long as Britain remains tied to the Common Market, no real improvement in agriculture can be expected. Only with self-government in Scotland can farmers hope to state their case effectively and                                                  51 Ironically, during the 2014 independence referendum campaign, the SNP’s leader, Alec Salmond pointed to the rise of Euroskepticism in the rest of the UK. He used this to argue that Scotland’s only chance of staying in the European Union was to vote for independence. (BBC News, 2014). 83 help to avoid the unnecessary insecurity and uncertain profitability of farming in Scotland in the future.” SNP, October 1974, Leaflet 5, sentences 7-8  In the 1979 election, the EEC was again the focus of SNP attacks. The following quote is representative of Scotland problems that were attributed to Europe: “Common Market? SNP told you so! Fishing grounds pillaged. Butter mountains. Senseless restriction on lorries. Prices rocketing. Bureaucracy gone mad.” SNP, 1979, Leaflet 2, sentences 1-7 In this election, though, arguments about the EEC are not linked to independence or devolution. As with the oil revenues debate, in 1979 the SNP argues for a policy that could be implemented without leaving the UK: “The SNP will take a constructive stand over Europe. We seek a looser European framework.” SNP, 1979, Leaflet 2, sentence 13-14   So, on the two substantive issues most affecting 1970s Scotland, we see SNP rhetoric shift from anti-system claims to pro-system claims. Moving to quantitative analysis, Table 3.2, below, shows the campaigns as a whole. There is a statistically significant shift between the two campaigns, in the predicted direction.52 When devolution was on the agenda, three-quarters of the SNP’s campaign material rejected the Westminster political system. This fits with the theoretical model, as all political agents knew that devolution would require a referendum victory. So, in the 1974 election, the SNP was not simply campaigning for seats at Westminster, it was laying the ground for a referendum campaign                                                  52 The chi-square statistic for the difference in the distributions of rhetoric between the two campaigns is 57.3020 with a p value of 0.000. 84 in the near future. To do this, it had to persuade voters first, that the current political system did not serve Scotland’s best interests and, second, that some degree of self-government was viable.53     Five years later, after the failure of the devolution referendum, the pattern is different. As Table 3.2 shows, a clear majority of the SNP’s leaflets were devoted to claims which accepted the constitutional system as it was. Interestingly, despite the SNP’s anger over the way the referendum had been conducted,54 the 1979 campaign is not simply devoted to grievances. Table 5 in Appendix 3 shows that, of the pro-system claims, approximately half attacked the Labour government while the other half claimed credit for concessions the SNP had gained at Westminster. This might be indicative of the SNP’s historically strong position in the 1974-9 Parliament: it had the highest number of MPs it had ever had55 and it was pivotal. This provides further support for the argument that in 1979, the SNP was appealing to voters on the basis of what it could achieve for Scotland at Westminster, rather than through Scottish self-government.   In terms of the strength of the association, there is a moderate relationship between the October 1974 campaign and anti-system messages on the one hand, and the 1979 election and pro-system messages on the other. The adjusted phi statistic for the relationship is 0.566, and the odds of a sentence from 1979 leaflets being pro system is 5.6 times that of a sentence from the October 1974 leaflets.56                                                    53 Appendix 3 includes break-downs of rhetoric for each election. Table 4 shows that in October 1974 the SNP addressed both concerns, laying out the benefits of self-government and their grievances against the status quo. However, the former were more prevalent – so although the October 1974 leaflets were anti-system they were not overwhelmingly negative in tone.  54 Chapter 5 examines the SNP’s involvement in the parliamentary passage of the Scotland Act 1978 and the 1979 no-confidence vote against the Labour government. 55 The 1975-79 Parliament remained the high water mark for the SNP until the 2015 UK election, when the party won 56 seats (BBC, 2015). 56 Conversely, the odds of a sentence from the Oct 1974 campaign being pro- system is 0.18 times that of a sentence from the 1979 campaign, when devolution was off the agenda.   85 Table 3.2 Categorizing campaign messages: SNP, October 1974 versus 1979  Oct 1974 Self-government ON agenda 1979 Self-government OFF agenda Pro-system sentences (Credit-claiming and Anti-government grieving)  27 sentences (24.5% of campaign) 284 sentences (64.5% of campaign)  Anti-system sentences (Positive case for Separation and Anti-system grieving) 83 sentences (75.5% of campaign) 156 sentences (35.5% of campaign) Total for campaign 110 sentences  440 sentences  N = 550 sentences Missing cases (sentences coded as neutral): 124   So, the quantitative evidence provides support for the hypothesis that the SNP would switch to a more pro-system approach after the failure of the devolution referendum. This shift is also borne out when looking at two particular problems claimed by the SNP: the squandering of oil revenues and farmers being driven out of business by EEC regulations. When devolution is on the agenda, the SNP poses self-government as the best solution to both these issues. Once it is off the agenda, the SNP instead proposes substantive policy changes that could be implemented by the Westminster government. This echoes the switch we saw in the IPP’s behaviour 80 years previously. The same substantive problems are raised, but an anti-system solution is offered when self-government is on the agenda, whereas a pro-system solution is favoured when self-government is off the agenda.   3.4 Results from scenario 3: a non-pivotal separatist party competing against the opposition This section focuses on the UK General Elections in February 1974 and 1983. Going into both these campaigns, the Conservative party had a majority government. In this period, the Conservatives had more Scottish seats than the SNP, but were far behind Labour in this region. This is true of the 86 elections we are examining and the previous elections (1970 and 1979 respectively). In both 1970 and 1979, all of the SNP-held seats had the Labour party coming in second (see Appendix 1, Table 1). Labour was therefore the SNP’s greatest threat, across Scotland and in specific seats. Commentators on Scottish electoral behaviour confirm this analysis. Bennie argues that from the 1960s onwards the Conservatives had lost their previous coalition of support in Scotland, and could not build a new one, so they instead focused on English seats (1997, 20-21).   February 1974 is an election when self-government is partially on the agenda. The Kilbrandon Commission had published its report 4 months prior to the election, recommending a Scottish Parliament. However, at the time of the election, no major party had endorsed these proposals – even the SNP was skeptical as it wanted full independence (Mitchell 2009, 114-118). Self-government was certainly a topical issue during the election, discussed by everyone, but at that point it was not clear that there would be a referendum on the subject. The SNP would therefore have a reasonable expectation that for the short-term they would only be running for Westminster elections. As explained in section 3.3, a referendum was eventually held in 1979 and devolution failed to meet the required vote threshold. Once the Thatcher government won the 1979 election, it repealed the Scotland Act, which had provided for devolution and the referendum. Therefore by 1983, self-government of any kind was a long way off.   As in the 1970s, oil was again a major campaign issue. In 1983, this issue was linked to the economic pressures that Scotland was facing. There was a UK-wide recession and the Conservative government had ended the “regional policy” which had funneled money to the UK’s deprived regions (Thatcher, 1993, 627). The government also opposed intervening to support failing industries, such as Scottish steel. Both Labour and the SNP campaigned on the basis that oil proceeds should have been used to 87 protect Scottish interests. Labour in particular decided to “play the Scottish card” by arguing that the Conservatives were a predominantly English party that had abandoned Scotland (Bennie, 1997, 37-38).   Table 3.3, below, shows that, contrary to predictions, both SNP campaigns are strongly anti-system. More than three-quarters of the campaign material in each year is devoted to attacks on the political system or to advocacy of self-government. There is a slight increase in pro-system messages in 1983, when self-government was off the agenda. The odds of an “off-agenda” (1983) campaign sentence being pro-system is 1.5 times that of a sentence from the “on-agenda” (February 1974) campaign.57  However, this difference is not statistically significant.58  Furthermore, the phi adjusted measure of association between the February 1974 campaign and anti-system messages (versus the 1983 campaign and pro-system messages) is only 0.102, which is classed by Bohrnstedt and Knoke as “virtually no relationship” (1982).   This is not explained by self-government only being partially on the agenda in the first election. If that was having an effect, both elections would be towards the pro-system end of the spectrum. Table 3.3 shows the opposite is the case: it is in fact the 1983 campaign which does not fit with predictions, being more anti-system than expected.  There are two possible explanations, both of which would fit with the theoretical model in Chapter 2. First, the SNP was not pivotal before either of these elections and its parliamentary numbers were negligible. In February 1974 the party was still in its electoral infancy, having only won its first seat in a 1967 by-election. Going into this election, it still only had 1 MP. At the 1979 election, it lost nearly                                                  57 Conversely, the odds of a sentence from the “on agenda” (Feb 1974) campaign being pro-system was 0.69 times that of a sentence from the “off agenda” (1983) campaign. 58 The chi-square for the difference in distribution of campaign messages between February 1974 and 1983 is 1.6609, with a p value of 0.197. 88 Table 3.3 Categorizing campaign messages: SNP, February 1974 versus 1983  Feb 1974 Self-government ON agenda 1983 Self-government OFF agenda Pro-system sentences (Credit-claiming and Anti-government grieving)   38 sentences (17.2% of campaign) 25 sentences (23.1% of campaign)  Anti-system sentences (Positive case for Separation and Anti-system grieving) 183 sentences (82.8% of campaign) 83 sentences (76.9% of campaign) Total for campaign 221 sentences  108 sentences  N = 329 sentences Missing cases (sentences coded as neutral): 121  all of its seats, meaning that it went into the 1983 election with only 2 MPs (see Appendix A, Table A.1). The SNP therefore had extremely limited resources with which to make any impact on legislation or even the debate at Westminster. This would make credit-claiming very difficult as it would simply not be credible to argue that the party was able to put pressure on the Government of the day in order to extract concessions. Sure enough, in both the February 1974 and the 1983 sets of leaflets, credit-claiming was used sparingly: only 11% and 8% of all sentences respectively (see Appendix C, Tables C.3, and C.6, respectively).   The SNP’s lack of legislative power would be conducive to anti-system claims, as the party is unable to have much impact. However, this motivation is clearly different to that of a period where the party is looking ahead to separate Scottish institutions and therefore less interested in British ones. One argument from the 1983 election campaign supports the former line of thinking, as it suggests the party would achieve more in the political system if it had more MPs: 89 “Scotland needs the SNP- A strong body of Nationalist MPs will make any Westminster Government think twice about giving Scotland a raw deal.” SNP, 1983, Leaflet 7, sentence 4.  Further evidence of this interpretation can be found by comparing the 1983 campaign to the 1979 campaign (discussed in section 3.3). Both of these elections were after the disappointment of the 1979 referendum, so self-government was off the agenda. However, in 1979, the SNP had its highest ever number of MPs (11) and was pivotal. We see a much higher proportion of credit-claiming sentences: 31% of all sentences, compared to 8% in 1983 (see Appendix C, Tables C.5 and C.6, respectively).  The second explanation also concerns the SNP’s opportunities to gain concessions for Scotland during the 1979-1983 Parliament. The SNP had limited leverage but the government of the day also had very little motive to increase spending in any region of the UK. My model points to one reason a government might lavish goods and services on a would-be separatist region: to undermine the mainstream opposition party (if the latter is strong in the region and the government is weak). However, there are, of course, many other factors controlling government spending, besides such electoral calculations.  For much of the 1979-1983 Parliament, the UK suffered a severe recession, curtailing government revenues and therefore expenditure. The Thatcher government also had its own ideological reasons for not wanting to use the credit-claiming strategy (or create the conditions for anyone else to do so). As Thatcher makes clear in her autobiography, she was suspicious of any regionally-based bureaucracy that measured its success in terms of increased spending: “The pride of the Scottish Office – whose very structure added a layer of bureaucracy standing in the way of reforms which were paying such dividends in England – was that public expenditure per head in Scotland was far higher than in England”  Thatcher, 1993, 627  90 Furthermore, the main opposition Labour party was electorally weak and internally divided at the time of the 1983 election, and therefore less of a threat. 59 So the lack of government expenditure in Scotland might have helped Labour there, but it was unlikely to win enough seats elsewhere in the UK to unseat the government.  I therefore argue that in 1983 the government was not acting in accordance with the model’s predictions, due to exogenous factors. What is most important, though, is how the model would predict that the separatist party would react to such circumstances. If the government has not in fact spent money in the region, then the separatist party cannot make any credible appeal to benefit-seeking voters. As Labour was both the strongest party in Scotland and the Official Opposition, we would expect it to attract the protest voters who dislike the incumbent government and its policies. This left the SNP with only its core group of voters, who are motivated primarily by separatism: i.e. they want some form of self-government. From this perspective, it is not surprising the 1983 campaign focuses on anti-system arguments almost as much as the February 1974 one, when the Kilbrandon Commission had just recommended devolution. The SNP could simply have been responding to electoral imperatives, rather than preparing for short-term exit.   This analysis is backed up by two sets of arguments deployed by the SNP in 1983. First when the party does raise grievances against another party (as opposed to the whole system), it is against the SNP’s chief electoral rival: “Labour and the Social Democrats [a new, centre-left party which split from Labour in 1981] have won By-Elections in Scotland this year . . . has it made any difference at all for Scotland’s prospects?? NO!!! The truth is that London only takes notice when Scots vote SNP.” SNP, 1983, Leaflet 3, sentences 22-23. Emphasis in the original.                                                       59 See, for example Sergeant 2001, 220-228, and 233-235; Hattersley, 1997, 277-294 and Clark, 1993, 5-8 on the state of party competition in this period.  91 On the face of it, this is a puzzling claim – in a primarily two-party system where the government has a secure majority, we would not expect the opposition party to have more leverage just because it won a couple of by-elections. What the SNP is arguing, however, is that it, not Labour, is best placed to be Scotland’s champion at Westminster. This is the same logic as the above quote from 1983’s Leaflet 7, which says the SNP can do more for Scotland when it has more MPs. This line of campaign argument points to Labour being the SNP’s chief electoral rival and therefore the party that must be discredited. It also suggests a willingness to continue engaging in the British political system, which reflects the circumstances of 1983, when any form of self-government looked unlikely.  The second set of arguments can be found when looking at the anti-system claims made in 1983. One example is: “Recent closures of Linwood Car Plant, Corpach Pulp Mill and Invergordon Smelter added massively to Scotland’s unemployment figures . . . and the Westminster government did nothing about it!! When Scotland votes SNP no London Government dares act this way – SNP votes are the ONLY real guarantee of any protection for Scotland. SNP, 1983, Leaflet 3, sentences 19-20.   This claim is a grievance against the overall political system because it makes negative descriptive claims (the closure of businesses and the resulting unemployment) and attributes these to the “Westminster government”, not the political party in power. However, the argument is ambivalent as it is advocating for a substantive policy which could take place within the existing political system: government intervention to support failing industries. Furthermore, this is a policy which is the clear opposite to that of the Conservative government, but would find sympathy from the Labour party of 1983. The fact that the SNP shies away from making a partisan point here could be due to its general disdain for all “London-based” parties, but it would also be consistent with its electoral situation: needing to distinguish itself from Labour, and not seeing the Conservatives as a threat. There is also a 92 suggestion that the SNP would engage with the political system under different circumstances: if it had more MPs, echoing the points made above about its tiny size in the 1979-1983 Parliament.   To sum up, comparing these scenario 3 elections does not provide clear support for the model, as the SNP remains very critical of the political system, even after devolution is off the agenda. However, it behaves as expected in the February 1974 election, when authoritative proposals for devolution have been made by the Kilbrandon Commission. The anti-system campaign of 1983 is traceable to the two factors which I argue affect separatist behaviour when self-government is off the agenda: legislative status and electoral competition. The SNP’s rhetoric in 1983 is what would be predicted for a non-pivotal party, ceteris paribus. It is not what would be predicted for an election where it was competing against the opposition party. However, this is actually a failure of the model to predict the government’s behaviour. Given that the Conservative government took a low-spending route, the SNP actually reacted as the model would predict in this scenario, by focusing on its core separatist vote. The hints that it might engage with the British system in the future suggest that this was as much about electoral calculation as it was about an expectation of early exit from the UK.   3.5 Results for scenario 4: a non-pivotal separatist party competing against the government For this final scenario, there are two sets of elections which fit the criteria. First, I look at the SNP in the final election for both the Blair and Brown governments (2005 and 2010 respectively). Second, I look at the Bloc Québécois’ first two elections, in 1993 and 1997.  3.5.1 The SNP in the Blair and Brown era In the mid-2000s, the SNP’s chief electoral rival was still the Labour party. The SNP did not have pivotal status, as the Labour government had a secure parliamentary majority before both the 2005 and 93 the 2010 elections. The important difference was at the sub-national level. The Blair government won a referendum on Scottish (and Welsh) devolution, and established the Scottish Parliament in 1999. For the first eight years of its life, the Scottish executive was controlled by a coalition of Scottish Labour and the Scottish Liberal Democrats, whose focus was making the existing settlement work rather than seeking further powers for the Parliament.60  As far as the Westminster Labour government was concerned, devolution was meant to end calls for independence. George Robertson, the first Scottish Secretary appointed to the Blair government, said that devolution would kill nationalism “stone dead” (quoted in Curtice and Seyd, 2009, 2). So, at the time of the 2005 UK-wide election, there were no proposals from either the British or Scottish government for further constitutional change.   This assessment is endorsed by scholars of devolution and federalism. While comparing the Scottish and Catalan sub-national governments, Greer argues that secession in either case “would take a social and political earthquake” (2007, 183).61  Cairney has the advantage of hindsight, as he is writing in 2011, after the SNP had won two Scottish Parliament elections. However, he concurs about the state of play in 2005, asserting you can divide Scottish devolution into two halves – from 1999-2005 and 2005-11. In the first half, he states that discussions of independence or further devolution were infrequent (2011, 220-221).   At the 2007 Scottish election, the SNP won the most seats for the first time, establishing a minority government. Although it did not have enough votes to pass its legislation on an independence referendum, the party did use its position to promote self-government. It launched the “National                                                  60 The Scottish First Minister, Jack McConnell, addressed this subject in 2002: “I don't believe we should be arguing about the powers of the Scottish Parliament for the next five or six years. We should get on with the business of delivering improved public services” and again in 2004: “Now is not the time to go into consideration of changing the powers of the Scottish Parliament.” Both quotes are reproduced in Cairney, 2011, 224. 61 This assessment was made before the SNP’s victory in the 2007 Scottish elections.   94 Conversation”, a consultation process on an independent Scotland’s future political system and economic policy. This culminated in a White Paper endorsing a referendum on independence (Scottish Government, 2009). Also important was the reaction of UK-wide parties. Labour, the Liberal Democrats and the Conservatives supported the alternative Calman Commission that evaluated the first decade of Scottish devolution. The Commission rejected independence as a viable option but supported the devolution of additional jurisdictions and “fiscal autonomy” for Scotland.62  This in turn was endorsed by the Gordon Brown’s government at Westminster.63  So, going into the 2010 election, the situation had reversed. Both levels of government now supported further constitutional change. Although they disagreed about whether Scotland should remain part of the UK, the overall subject of self-government was back on the agenda.   Table 3.4, below, shows the relationship between an on-agenda election and pro-system messages. As predicted, we do see a drop in pro-system messages in the 2010 election, which was a few months after the SNP’s “National Conversation” on independence. The difference between the two elections is statistically significant at a 90% confidence level.64  The odds of the “off-agenda” campaign (2005) having pro-system messages is 6.7 times of the “on-agenda” campaign (2010).65                                                   62 This meant reducing income tax thresholds for Scottish residents by 10 percentage points, with a corresponding reduction in the block grant from the Westminster Parliament to the Scottish Parliament. The Scottish Parliament could then choose to use this “tax room” by setting Scottish income taxes in addition to those collected centrally. The Scottish Parliament would also have control over the tax levied on some transactions, such as house sales and air tickets, and there would be fewer restrictions on its ability to borrow money. See Commission on Scottish Devolution, “Serving Scotland Better: Scotland and the United Kingdom in the 21st century”, 2009, pp 89-110. 63 See UK Government, “Scotland’s Future in the UK”, Command Paper 7738, 2009.   64 The chi-square statistic for the difference in the distribution of campaign messages between 2005 and 2010 is 3.7091, with a p value of 0.054. 65 Conversely, the odds of a sentence from the “on agenda” campaign (2010) being pro system is 0.15 times that of a sentence from the “off agenda” campaign (2005).   95      Table 3.4 Categorizing campaign messages: SNP, 2005 versus 2010  2010 Self-government ON agenda 2005 Self-government OFF agenda Pro-system sentences (Credit-claiming and Anti-government grieving)  24 sentences (70.6% of campaign) 16 sentences (94.1% of campaign)  Anti-system sentences (Positive case for Separation and Anti-system grieving) 10 sentences (29.4% of campaign) 1 sentence (5.9% of campaign) Total for campaign 34 sentences  17 sentences       N = 51 sentences.      Missing cases (sentences coded as neutral): 2  However, the phi-adjusted measure of association is only 0.364, classed by Bohrnstedt and Knoke as a “weak” relationship (1982).  As the table shows, the SNP’s 2010 campaign was still largely pro-system, though to a lesser degree than in 2005. This ambivalence can be seen in the way the SNP explains its role in the Westminster Parliament to potential voters. At the beginning of the leader’s foreword, in both years, the SNP uses Scottish identity as a rallying point. However, this is linked to the importance of Scottish representation in the existing, UK-wide institutions: “The more SNP MPs elected to Westminster, the more we can do for our country.” SNP Manifesto, 2005, sentence 3  “This election is about ensuring a strong team of Scottish MPs in the House of Commons – a team of SNP MPs. Local champions, who will be there working hard for you, your community and Scotland.” SNP Manifesto 2010, sentences 1-2  96 So, the first thing the reader sees, in both cases, is a reaffirmation of the importance of the Westminster Parliament.66 The SNP is therefore cast as the logical choice for the benefits-seeking voter. This is someone who does think of Scotland as a distinct entity, but is primarily concerned with getting a better deal for Scotland from the rest of the country, not with Scotland governing itself. Both manifestos do refer to the Scottish Parliament, but later in the foreword (in sentence 10 in 2005 and sentence 8 in 2010). On policy, both manifestos are again similar, as they list a range of substantive policies which could be implemented without any constitutional change. Both years call for an increase in the number of jobs, the size of pensions and the availability of health services. The 2005 manifesto also promises increased policing and changes to municipal taxes (SNP, 2005, sentences 13-14; SNP 2010, sentence 11).   However, one clear difference fits with my predictions. There are more mentions of independence in 2010 and this is more closely integrated to everything else that is said. In 2005, there is only one sentence giving an anti-system message in the whole foreword. It come two-thirds of the way through, following an explanation of how the SNP will represent Scots at Westminster:  “And we will always argue the case for Scottish Independence.” SNP, Manifesto 2005, sentence 11  This statement is followed by a discussion of the substantive policies that the SNP favours of the type noted above. So, although the SNP affirms its support for independence, this is not linked logically to anything else in the foreword. It is not offered as the solution to any contemporary Scottish problems, nor are grievances raised against the rest of the country or the existing political system. Furthermore,                                                  66 This echoes the message from 1983, on the need for more SNP Members of Parliament – see section 3.4. 97 the foreword does not put forward any positive argument for independence – there is no statement about any benefits that self-government could bring. This, then, looks like a reaffirmation of the policy to those who are already committed to it: the SNP’s genuinely separatist core voters. However, in this foreword, the party is not trying to convince anyone of the merits of independence, which we would expect if there was an independence referendum in the offing. The rest of the foreword seems instead to be an appeal to benefits-seeking voters.  In 2010, independence is cast as the central organizing idea of the manifesto: “Running through this manifesto is the SNP vision of a new future for Scotland, independent, socially just and economically secure.” SNP Manifesto, 2010, sentence 32  This manifesto is setting out a case for independence, seeking to persuade those who are not already committed supporters. In the above quote, we have the positive case for separatism, pointing to what the party argues are the moral and material benefits of self-government. We also see the negative case, when the SNP attacks all other political parties, defining them as “London-based”. We see this in a general statement at the beginning of the foreword, where being centred in the UK’s capital is equated with being contrary to Scotland’s best interests: “The London parties all offer the same thing – the wrong priorities for our nation.” SNP Manifesto, 2010, sentence 3. This argument is later applied to the specific situation of the global financial crisis and the UK recession, which had lasted since 2008:  “We are working hard for economic recovery and new jobs and opportunities for families and communities. However the London parties’ proposed cuts pose a threat to this recovery.”  SNP Manifesto, 2010, sentences 20-21 98 Here, the positive claim that unemployment could fall in the future is juxtaposed with the negative claim that all other parties are jeopardizing Scotland’s recovery. This, then, is not simply anti-government grieving which could be addressed by replacing the Labour party in government. Rather, it is a claim that no UK-wide party can be trusted – their policies will hurt Scotland’s economy. This fits with the following quote; taken together the message is that Scotland is not well served by either the UK’s political parties or its institutions: “The real alternative to the discredited Westminster system is a fresh, independent future for our nation.” SNP Manifesto, 2010, sentence 33   Finally, the anti-system message also appears when the SNP addresses the paradox of a separatist party running for seats in the central legislature. First, it makes a temporal argument, indicating that the current document has a short-term focus (through the use of the word “now”), but reiterating that its medium and long-term focus is still independence: “As a party and as Scotland’s government, our focus is first and foremost on winning the best deal for Scotland. We are ambitious for you and your family and confident in Scotland’s ability to succeed, now and with independence. ” SNP Manifesto, 2010, sentence 17  Second, the party argues that it can use its presence at Westminster to gain self-government, intimating that it will put pressure on the incoming government to grant concessions: “Our MPs will work to ensure that Scots get the opportunity to be consulted on our own constitutional future. Incredibly, the Labour Party now propose an early referendum on the Alternative Vote and the House of Lords while they seek to block the Scottish Parliament holding a referendum on Independence.” SNP Manifesto, 2010, sentences 30-31  To sum up, in both the 2005 and 2010 election, the majority of the leader’s foreword was devoted to messages which supported the existing political system. The drop in these messages in 2010 was not 99 quantitatively substantial. Both years used the “local champion” argument that will appeal most to benefit-seeking voters who might identify as Scottish, but want to stay within the UK. However, I argue that when it was used, the anti-system message was deployed differently in the two elections. In 2005, the party simply restates its commitment to independence, but does not build any case for it. This might reassure voters who are already separatist, but does not seem calculated to change anyone’s views. In 2010, the explicitly separatist message is linked to different parts of the manifesto: achieving independence is a reason to elect SNP MPs to Westminster, it is also the prescription for Scotland’s current problems and it is the route to a positive future.   3.5.2 The BQ in the Mulroney and Chrétien era. The Bloc Québécois was formed in 1990 by a small group of Canadian MPs who had defected from both the government and official opposition federal parties. The 1993 and 1997 elections were the first two federal contests in which its candidates ran for office. Its electoral and legislative position was the same as the SNP in 2005 and 2010. First, the government of the day had a secure parliamentary majority going into both elections: 27 seats in 1993 and 61 in 1997 (see Appendix 1). The BQ was therefore not pivotal.   The BQ’s primary rival in Québec was the government party in both elections. Interestingly, though, this government party was not the same, as the electoral situation in Québec had shifted so dramatically in the 1993 election.67  In 1988, the Progressive Conservatives (PC) won both a majority in the Canadian House of Commons and of the Québec seats (84% - see Appendix 1). The BQ of course won no seats as it did not yet exist. Therefore, going into the 1993 election, the PC government party was                                                  67 There are many accounts of the 1993 Canadian election and the shift from a 3-party, pan-Canadian system to a 5-party, regionally-fragmented system. See, for example, Gidengil, 2012; Marland et al, 2012; Cardinal, 2005; Carty, Cross and Young, 2000; Nevitte et al, 1999 and Young, 1999. 100 the one with the most seats to target. At that election, the PC party was reduced to only 2 seats across all of Canada. The BQ took its place as the dominant Québec party, with 72% of the province’s seats. The Liberal Party of Canada formed a new majority government, and was also the second party in Québec, with 25% of the province’s seats. By the time of the 1997 election, therefore, the Liberal government was the primary electoral rival to the Bloc. So, in both elections, the Bloc was primarily trying to win seats or defend them from the party in government.   The fact that there was a change in the government party allows a test of the significance of the left-right spectrum. Chapter 2 argues that a separatist party will try to draw support from across this spectrum. The Bloc’s own history supports this, as its founders had defected from both the Progressive Conservatives and the Liberal Party in order to mobilize around their common interest in Québec nationalism (Cardinal, 2005; 29-33). If my argument is correct, a separatist party should not attack a government on the basis of how left-wing or right-wing it is. Rather, separatist claims about the government depend on two factors: whether the government will make concessions and whether the government is an electoral threat to the separatist party.   For my theory, the crucial difference between the two elections is that they occurred on either side of the 1995 referendum on Québec sovereignty. The constitutional status of Québec had been on the political agenda since 1982. In that year, amendments to the Canadian constitution were passed without the consent of the Québec provincial government. When the Progressive Conservatives came to power in 1984, Brian Mulroney devoted considerable time to negotiating a package of constitutional amendments that would be acceptable to both Québec and the other Canadian provinces. He was unsuccessful, triggering the defection of some of his Québécois Cabinet ministers, in order to form the Bloc.  101 After the failure of a second constitutional package, the Québec provincial government established the Bélanger-Campeau Commission. Its mandate was to examine Québec’s future options for achieving more autonomy and a distinct status from the rest of Canada. Reporting in 1991, the Commission set out two routes: seeking the repatriation of as many jurisdictions as possible from Ottawa to Québec City, or holding a referendum on political sovereignty (Cardinal, 2005, 31-32; Landry 1991). The provincial Québec government at that time was not separatist, but it set up two parliamentary committees to consider both tracks. This is the basis on which I categorize 1993 as an election where separation was “partially” on the agenda. There had been several authoritative sets of proposals, from both the federal and provincial governments which proposed greater autonomy for Québec. The last of these, the Bélanger-Campeau Commission, had openly discussed sovereignty. However, in 1993, neither level of government was endorsing sovereignty.   In 1994, a sovereigntist government was elected in Québec. It immediately started making preparations for a referendum on political and economic sovereignty for Québec (Young, 3 and 13; 1999). Sovereignty was narrowly defeated in the October 1995 referendum, by a margin of 1.4% (CBC, 1995). Many commentators have argued that the close result meant that the sovereigntist movement, and the federal government’s reaction to it were key election issues in 1997 (see, for example, Cardinal 2005, 50-51; Nevitte et al, 1999, 17 and 117-8 and Young, 1999, 18, 74 and 83). I agree that the issue was frequently discussed during the 1997 federal campaign. However, as set out in Section 3.1, the test for whether self-government is “on the agenda” asks whether the separatist party can reasonably expect to exit the political system within the next electoral cycle. In 1997, this was not likely within the next 4-5 years. The sovereigntist recommendations of the Bélanger-Campeau Commission ran their course when the 1995 referendum was held. Any further discussion by either level of government was focused on the process for holding a referendum, or federal expenditure designed to keep Québec within the 102 federation. Neither government was discussing what a sovereigntist political system would look like, nor the economic position of an independent Québec. Lucien Bouchard, the new sovereigntist Premier of Québec (and former leader of the BQ) had ruled out another referendum until there had been another provincial election (Young, 1999, 88). Therefore, sovereignty was not an immediate option for the Bloc at this point: it would have to sit in the federal Parliament for at least one more legislative and electoral cycle.  Table 3.5, below, compares the rhetoric in the 1993 and 1997 election campaigns. It shows a clear shift from anti-system messages in 1993 to pro-system messages in 1997, which fits with the model’s predictions.68  The phi adjusted measure of association between the campaign and the proportion of manifesto sentences devoted to anti-system rhetoric is 0.699, which Bohrnstedt and Knoke term a “moderate” relationship (1982). Put another way, a sentence in the BQ’s 1997 manifesto, after the failure of the sovereignty referendum, had 25 times the odds of being pro-system than a sentence in the BQ’s 1993 manifesto, before the referendum was held.69  So, the overall patterns of rhetoric in this case fit with my argument that once self-government is no longer on the agenda, a separatist party will shift electoral strategies. As it has to remain in the current political system for the foreseeable future, it begins to consider how it can hold on to parliamentary seats, for example by campaigning for more expenditure in its region. It may also start to sound more like a mainstream opposition party, attacking the government of the day on substantive policy, rather than the legitimacy of the whole system.                                                    68 The difference in the distributions of the two campaigns has a chi-square statistic of 39.5418, with a p value of 0.000. 69 Conversely, the odds of a sentence in the 1993 manifesto (when sovereignty was on the agenda) supporting the Canadian political system were 0.04 times that of a sentence in the 1997 manifesto. 103 Table 3.5 Categorizing campaign messages: BQ, 1993 versus 1997  1993 Self-government ON agenda 1997 Self-government OFF agenda Pro-system sentences (Credit-claiming and Anti-government grieving)  9 sentences (25.7% of campaign) 52 sentences (89.7% of campaign)  Anti-system sentences (Positive case for Separation and Anti-system grieving) 26 sentences (74.3% of campaign) 6 sentences (10.3% of campaign) Total for campaign 35 sentences  58 sentences    N = 93 sentences   Missing cases (sentences coded as neutral): 15   In 1991, when the Bloc had just been formed, and was setting out a “pre-election” manifesto, it refers to Québec’s linguistic and cultural heritage, and endorses the provincial-level political institutions in order to build a case for sovereignty: “Our natural allegiance is Québécois. Our home territory is Québec home to a people dedicated towards culture and the French language for whom we intend to seek sovereignty. We consider the National Assembly of Québec, as the legal and de facto supreme democratic institution of the Québécois people. This is where they [the Québécois] should exercise their sovereign authority.” Bloc Québécois pre-election manifesto, 1991, sentences 6-9  By 1997, after the failure of the referendum, there is still a positive tone. However, the important shift is that much of the manifesto is now devoted to setting out substantive policies which will improve life for its constituents. So, positive descriptive claims (such as the ones in the 1991 quote) are no longer being tied to sovereignty. Rather, these are policies that the Bloc will pursue as a federal political party sitting in the Canadian Parliament. The party explains its role as: “The only defender of the rights of Québec in Ottawa.” Bloc Québécois manifesto, 1997, sentence 7. 104 So, here it is implicitly accepting the constitutional status quo, at least for the short-term. Its reference to “rights” can appeal to both those concerned about the provincial government’s autonomy within its own jurisdictions (who are more likely to be separatist voters) and those who simply want Québec to get goods and services from the federal government (pro-system, benefits-seeking voters).   The substantive policy area which is mentioned first in the 1997 manifesto is the economic situation, in particular the federal government’s efforts to reduce the deficit. Looking forward to the expected federal surpluses, the Bloc declares that it had a plan for spending the money (1997 manifesto, sentence 9). The reforms it then lists all pre-suppose the existing federal system, concentrating on tax cuts for lower-income families, reforms to Employment Insurance and federal government-funded parental leave. For the latter program, the Bloc insists that it will be managed by the Québec provincial government. However, even this assertion of provincial rights pre-supposes a continuing federal constitutional structure – it shows that the Bloc expects the Québec provincial government to continue receiving fiscal transfers from Ottawa for the foreseeable future.   By contrast, in the 1993 manifesto, there are no specific economic policies proposed. When the economy or Canada’s fiscal situation is mentioned, it is in very general terms. In this example, the party is talking about the Canadian system’s impact on Québec:  “Worse, a system that takes us along with its political and economic drift, with administrative waste being only one aspect of this.” Bloc Québécois manifesto, 1993, sentence 4.  This grievance about the whole political system is a set-up for the following paragraph of the manifesto, which says that Québec can either accept the situation or take “charge of its own destiny” 105 (Bloc Québécois manifesto, 1993, sentence 6). The latter option is explicitly stated to be sovereignty, as the party begins to explain what it is trying to achieve in Ottawa: “It is obviously not enough to say that sovereignty is necessary, we have to develop an effective strategy to reach it. Hence the second finding that led to the creation of this new party: Québec must refuel its political force in Québec City and in Ottawa.”  Bloc Québécois manifesto, 1993, sentences 7-8.  Therefore, references to the economy and to government expenditure are used very differently in the two campaigns. In 1993, when both levels of government were discussing new constitutional options for Québec, the subject is used for anti-system grieving, which segues into the case for sovereignty. In 1997, after the failure of the referendum, the Bloc puts forward specific economic and fiscal policies which are intended to be implemented within the existing constitutional framework. The shift in rhetoric about economic policy and public expenditure therefore reflects the broader move from an anti-system manifesto to one that engages with substantive debates within the existing political institutions.   3.6 Conclusion This chapter examined the impact of the prospect of self-government, which I argue is the most important factor affecting separatist party behaviour. I compared elections where self-government was on the agenda to those where it was no and did so in four different electoral and legislative scenarios. The empirical results set out here provide support for the model in three of the four scenarios I examined (the exception being scenario 3, a non-pivotal party competing against an opposition party).   All of the parties studied shifted their electoral rhetoric over time. The Irish Parliamentary Party ignored the incentives it had to credit-claim in 1886, when Home Rule was a distinct possibility. In 1895, when Home Rule had receded, these incentives appeared much more important. Rather than 106 focusing on a self-governing Ireland, its candidates instead devoted time to explaining what concessions they got on land reform as a result of their pivotal status and their relatively friendly relationship with the Liberal government. The Irish case was an example of scenario 1: a “hard case” where the other two independent variables I examine were pushing the separatist party towards a pro-system, credit-claiming approach. Despite these incentives, the IPP sticks with anti-system grieving in 1886. This provides important support for my argument that the prospect of self-government will outweigh all other factors for separatist parties.   The rest of the chapter tested this argument in three other electoral and legislative scenarios. The parties’ behaviour fits with predictions for scenario 2, when they are pivotal and competing with the government (the SNP in the 1970s). The separatists also meet predictions in scenario 4, when they are not pivotal and are competing with the government (the BQ in the 1990s, and the SNP in the 2000s, though the evidence is quite weak in the latter case). The one situation that does not support my hypotheses is scenario 3, where the separatist party is not pivotal and is competing against the opposition party (the SNP in the February 1974 and 1983 elections).  Here, the party maintained a strongly anti-system message even after the failure of a devolution referendum. As argued in section 3.4, this may be a reflection of the fact that the odds were stacked against the SNP in this period: its parliamentary caucus was tiny, the government had a strong majority and the economic situation and government’s ideology meant that funding was being cut for Scotland, not increased. This meant that credit-claiming would not be credible. However, the other pro-system approach of anti-government grieving would also be electorally unwise: in a primarily two-party system this could simply drive voters to the Official Opposition party, which was the SNP’s electoral rival.   107 This last point fits with another part of my model: electoral calculations. The next chapter considers their impact on separatist behaviour, along with the final variable in the model: pivotal status. That chapter asks what separatist parties do when self-government is off the agenda. Does separatist behaviour start to resemble that of mainstream parties, and is it driven by either electoral or legislative factors?   108 Chapter 4: How do separatist parties appeal to voters when self-government is off the agenda? This chapter focuses on separatists’ electoral behaviour when self- government is off the political agenda. It asks how a party can remain relevant to voters when it has little chance of achieving its primary policy objective. I argue that when their primary goal is not an immediate prospect, separatists will instead focus on electoral competition in the current system. It is during these periods that my model of separatist party behaviour diverges most sharply from the predictions of the anti-system party literature reviewed in Chapter 1. These scholars argue that anti-system parties will, by their very nature, always attack the political regime, regardless of changing circumstances. However, the model in Chapter 2 predicts separatists will vary their rhetoric when self-government is off the agenda. When a separatist party competes against the governing party it will focus on grievances; when it competes against an opposition party, it will credit-claim. To test this argument, this chapter examines separatist behaviour in scenarios V-VIII from Figure 2.2 in Chapter 2. These are all cases where self-government is unlikely; however the second and third independent variables change – so the separatists’ electoral and legislative position varies.    I use text analysis of election manifestos and campaign leaflets to assess whether the campaign messages of the Scottish National Party (SNP), Irish Parliamentary Party (IPP) and Bloc Québécois (BQ) shift between elections. The chapter begins by setting out six pairs of election campaigns. These were chosen to isolate the effect of this chapter’s independent variables: the separatists’ status in the legislature and the separatists’ primary electoral competitor. This section also reviews how I analyse electoral rhetoric, using the same conceptual framework as in Chapter 3. However, in this chapter, the dependent variables are “grieving” and “credit-claiming”, whereas in Chapter 3, I was looking for “pro-system” and “anti-system” rhetoric. In the first of the results sections I examine the impact of electoral 109 competition, comparing elections where the separatists’ primary competitor was in government to those when the competitor was in opposition. In the second results section I test the prediction that legislative status does not affect separatists’ electoral rhetoric. This section compares separatists’ campaign behaviour in periods when they are pivotal to periods when they had no such leverage.  4.1 Case selection and data analysis 4.1.1 Selection of election campaigns Self-government was off the agenda during all the election campaigns examined in this chapter. This does not mean that the separatists’ region had no powers to govern itself. Rather, we are focusing on periods when there were no authoritative proposals to change the division of power between the central and sub-national governments (if the latter exist). So, during the electoral campaigns in this chapter there were no proposals for self-government, devolution of power or outright separation.70  The separatist party could anticipate participating in at least one more statewide election campaign and one more session of the central legislature: so it had to explain its role in the existing system to the electorate.  Figure 4.1 shows how the selected elections can be compared to test the impact of the first independent variable, electoral competitor. Each pair of cases includes an election where the separatist party competes against the government of the day and one where it competes against the mainstream opposition party. A shift in electoral competitor may be due to a change in government (i.e. the                                                   70 For the SNP this covers the period between the failure of the 1979 referendum on Scottish devolution and the report of the cross-party Constitutional Convention in 1996. It also covers the period after devolution began (in 1999) through to 2007 when the Scottish National Party won control of the Scottish Executive and used this position to push for a referendum on independence.  For the Irish Parliamentary Party, self-government was off the agenda whenever there was a Conservative government, as the latter was committed to maintaining a unitary state. It was also unlikely during the Liberal governments between 1906 and 1911, earlier because it was not a government priority and later (after 1909) because it would have been blocked by the House of Lords (Wicks, 2006). For the Bloc Québécois, sovereignty was off the agenda after the narrow “no” vote in the October 1995 referendum on sovereignty. 110    Separatists are competing with the Government Separatists are competing with the Opposition Separatists are not pivotal in the legislature   SNP, 2001 IPP 1906            Predicted behaviour: Grieving   SNP, 1983 IPP 1909-10  Predicted behaviour: Credit-claiming Separatists are pivotal in the legislature    SNP 1979: Aberdeen North constituency  Predicted behaviour: Grieving       SNP 1979: Moray and Nairn and       Aberdeenshire East constituencies  Predicted behaviour: Credit-claiming Figure 4.1 Paired elections, varying electoral competitor and keeping legislative status constant  separatists’ competitor lost an election and is now in opposition) or a change in competitor (i.e. there was a shake-up of the parties dominant in the region, with one of the mainstream parties gaining strength there).   In this set of cases, legislative status is held constant. The top row of Figure 4.1 shows a pair of cases where the separatist party was not pivotal in the Parliament immediately before the election. There are two pairs of elections which fit these parameters: the SNP in 2001 and 1983 and the IPP in 1906 and in 1909-10.71  My theory predicts that the change in electoral competitor will be sufficient by itself to change the separatists’ electoral rhetoric, from grieving to credit-claiming. The latter type of behaviour differs from that predicted by the literature on anti-system parties. So, testing this prediction is a crucial                                                  71 See Appendix 1 for more detail on the legislative and electoral conditions affecting these pairs of elections. 111 test for my theory. The upper right cell in Figure 4.1 is a particularly hard case for the model to predict, as the independent variables push the separatists in different directions. The party is not pivotal and so has limited leverage with which to extract concessions for its region. However, it is competing with an opposition party - and the latter is likely to appeal to protest voters unhappy with the current government. I argue that in this scenario, separatists are more likely to win over a different type of voter: those seeking benefits for their region. So, the separatists’ best electoral hope is a campaign based on credit-claiming, despite their non-pivotal legislative status.   The second row of Figure 4.1 makes the same comparison of elections: those with a government competitor versus those with an opposition competitor. However, in these cases the separatist parties were pivotal. The prediction is the same as it is for the cases in the first row. This is because my theoretical model states that a change in electoral competitor has the same impact, regardless of whether separatists are pivotal or not. Finding paired elections which fit these parameters is challenging. As will be shown in Figure 4.2, there are elections where the Bloc was pivotal and competing with the government party. However, there is no clear-cut comparator election.72    To counter this weakness, I include a constituency-level analysis for the SNP in 1979, using leaflets from the University of Aberdeen’s collection. Across the whole of Scotland, the SNP’s competitor was the governing Labour party. However, the Aberdeen collection includes two constituencies where the SNP MP’s closest electoral rival was the opposition party; one seat with a sitting government MP and                                                  72 The Québécois electorate is prone to sudden swings, so throughout the period when the Bloc was pivotal, all three of the pan-Canadian parties were plausible electoral rivals at different points in time. In the 2011 election, one of the opposition parties took most of the Bloc’s seats. However, what is important is the party leadership’s perception of electoral competition in the run- up to the electoral campaign, not the actual election results.  I therefore examined opinion poll data for the six months prior to the Canadian elections to help categorize each election. This showed that in the run-up to the 2011 election, there was no clear-cut rival for the Bloc. I therefore do not use 2011 as a test for this part of the hypothesis: we would not know if the results were due to the hypothesized causal factors, or due to mis-categorization of the case.  112 two with tight contests between UK-wide parties where the SNP was in 3rd or 4th place.73 These local-level contests allow us to observe if the SNP’s rhetoric changes with its electoral competitor, while holding pivotal status constant.  The second half of this chapter tests the prediction that a separatist party’s legislative status does not affect its electoral rhetoric. Figure 4.2 sets out the paired elections where the separatist party was pivotal and not pivotal, holding electoral competitor constant. The cases in the top row are elections where separatists were competing against the mainstream opposition. They can be thought of as another crucial test of my theory as the prediction of credit-claiming is different from that of the literature on anti-system parties. Here I compare the IPP’s behaviour in the 1909-10 and the December 1910 elections.74    Separatists are not pivotal in the legislature Separatists are pivotal in the legislature Separatists are competing with the Opposition   IPP 1909-10            Predicted behaviour: Credit-claiming   IPP Dec 1910  Predicted behaviour: Credit-claiming Separatists are competing with the Government   SNP 2001  BQ 2004   Predicted behaviour: Grieving   SNP 1979  BQ 2006 & 2008  Predicted behaviour: Grieving  Figure 4.2 Paired elections, varying legislative status and holding electoral competitor constant                                                  73 These categorizations of the electoral contest are true of both the 1974 and 1979 election results (Rallings and Thrasher, 2000). 74 See Appendix 1 for more details on these cases. 113 The final set of comparisons are on the bottom row of Figure 4.2. These cases vary pivotal status when the separatists are competing with the government of the day. Again, the theory predicts no change in behaviour: the need to attack the key electoral rival should mean the separatists grieve even when they have legislative leverage. There are two pairs of cases I can use to test this prediction. First, I compare the SNP in 2001, when it was not pivotal, to 1979 when it was. Second, I use the BQ in the period between 2004 and 2008, when there were three federal elections. Going into the 2004 election, the Bloc was not pivotal, whereas it was for both the 2006 and 2008 elections.  The opinion poll data noted in footnote 72 shows that the BQ’s key rival going into both the 2004 and 2006 elections was the governing Liberals whereas in 2008 it was the governing Conservatives. This shift in Québec’s electoral landscape allows us to test another possibility: that the incumbent government’s ideology makes a difference to the electoral rhetoric of separatists. In the 2006 and 2008 campaigns, the BQ had been pivotal and was competing against the incumbent government, so according to my argument we should see no change in its behaviour. Comparing both of these campaigns to the 2004 non-pivotal case allows us to check whether this still holds true even after a shake-up of the party system in the separatists’ region. Does the BQ campaign differently against a centrist rival than it does against a right-wing rival, or is the Bloc’s behaviour driven purely by electoral considerations?   Taken together, the elections set out in Figures 4.1 and 4.2 isolate each of the two independent variables. They cover cases which are both easy and hard for the theory to predict correctly. Finally, they provide crucial tests of the scenarios where the theory’s prediction diverges most sharply from that of existing literature: that a separatist party will use credit-claiming rhetoric when competing with the mainstream opposition party.   114 4.1.2 Measurement of electoral behaviour When parties communicate with potential voters during an election campaign, they are making descriptive and attributive claims. Descriptive claims are statements about the way the world is and attributive claims make causal arguments about who or what caused that state of affairs. In this chapter, I focus on descriptive and attributive claims in order to test whether separatist parties are structuring their election campaigns around grievances or credit-claiming.  Grievance-based rhetoric is based on a negative descriptive claim. If a separatist party makes descriptive claims which refer to negative impacts on its region, it is basing its electoral message on grievances. However, as noted in Chapter 3, this grieving may be directed at the choices made by the government of the day and the actions of individuals or it may be a fundamental challenge to the political system itself. The former is the same type of rhetoric that any opposition party might use, as its aim is for voters to replace the government or change the government’s behaviour, not seek constitutional change. Voters are told that their region is suffering harm and the solution is to send a separatist party to the central legislature that will be more adept at understanding and representing the region’s interests than another (countrywide) party that runs in the region. Anti-system grieving means that the party’s descriptive claims state that policies, actions or programs have caused harm to its region. Crucially, this is attributed to the political system as a whole, or all other political parties. The party therefore urges readers to vote for the separatist party in order to demonstrate support for self-government.   Credit-claiming consists of two elements. First the party makes descriptive claims that the region has received benefits. Second, the party attributes these benefits to its MPs’ efforts to represent the region in central institutions. So, credit-claiming is distinct from positive descriptive claims which are 115 attributed to self-government. The latter is simply the positive case for separatism, which we would expect from separatist parties. It obviously does not support the current political system in the way that credit-claiming does. Figure 4.3, below, sets out how descriptive and attributive claims combine in electoral rhetoric.  As in Chapter 3, I hand-coded the leaders’ forewords in manifestos from the selected election campaigns.75  Appendix 2 sets out the data sources, the coding procedure and the detailed data coding rules used, as well as the use of campaign leaflets for elections when manifestos were not available. The tone and length of manifestos varies immensely over my party cases: from the IPP’s 14-sentence manifesto in the 1909-1910 campaign to the BQ’s 206-page platform in the 2005-06 campaign. Furthermore, the Bloc’s manifestos are issued in French rather than English.76 However, I am focusing  on changes to each party’s rhetoric over time, so these cross-party variations in language, format and length should not undermine the validity of my findings.                                                         75 Manifestos for the SNP and BQ were downloaded from the Comparative Manifestos Project database, the University of Laval’s Electronic Political Texts database and the parties’ own websites. The IPP’s manifestos were found in contemporary news reports in The Times, which reproduced the manifestos verbatim. As the IPP’s manifestos were short and not broken into sections, entire manifestos were coded for this party. 76 The BQ also issued separate English-language platforms which are much shorter than the French version. In 2004 and 2006 the BQ’s leader wrote separate forewords for each of the English and French language platforms whereas in 2008 the English foreword was a direct translation of the French foreword. For 2004 and 2006, I used the French versions and translated them into English. The Bloc is primarily a Francophone party appealing to Francophone voters, so using the French-language platform captures the message that the Bloc is sending to the majority of its voters. I used Google Translate for the initial translation and then edited the resulting text into grammatical English. Thanks to Elise Leclerc-Gagné for checking the final English translation against the French original for accuracy.  116                                                       Descriptive claims:  positive  Credit-Claiming: Positive claims which are attributed to the separatist party’s actions within the current system or the current government               The Case for Self-Government Positive claims about the future which are attributed to increased self-government                              Anti-Government Grieving Negative claims which are attributed to the party in government, or specific politicians or policies.             Anti-System Grieving Negative claims which are attributed to the overall political system or country.  Figure 4.3 Categorizing campaign messages                                                                         Attributive claims:  not challenging  political system Attributive claims: challenging political system                  Descriptive claims: negative 117 4.2 Testing the impact of electoral competition This section discusses the results of the comparisons in Figure 4.1, which examine the differences between a separatist campaign when its electoral rival is in government and when its rival is in opposition. All the cases considered here hold legislative status constant.  4.2.1 When the separatist party is not pivotal 4.2.1.1 The SNP: 1983 v 2001 As a small party in an essentially two-party system, the SNP is the closest fit to the model set out in Chapter 2 of a separatist party trying to draw votes from either the government or the official opposition. As noted in Section 4.1, its key rival for Scottish seats has always been the Labour party. In the 1983 election, the Labour party had been in opposition for one parliamentary term whereas in 2001 it had been in government for the same period of time. In both of these periods the government of the day had a secure majority, so the SNP was not pivotal at Westminster, allowing us to rule out legislative status as a cause of any changes in its electoral rhetoric.  Table 4.1, below,  shows all the SNP’s descriptive claims from both elections, as well as showing who Scotland’s circumstances were attributed to. So, the top half of the table shows the institutions or groups that the SNP blamed for the problems that it highlighted in its campaign material. The bottom half of the table shows who is credited for any positive situations in Scotland.  Attributions were coded as blaming the political system if they referred to the government using institutional or geographical terms (i.e. “London”, the “London government” “British governments” “the UK” or “Westminster”) or if they explicitly blamed both mainstream parties. Conversely, whenever the government (or opposition) was blamed using the name of the party, or its leader, and  118 Table 4.1 Descriptive claims in SNP campaigns: who gets the blame and who gets the credit?77    1983  Opposition competitor; not pivotal 2001  Government competitor;  not pivotal Z-statistic for the difference between the two years (p value,1 tailed test) Negative descriptive claims attributed to: The political system: anti-system grieving  25 (71% of all negative claims) 2 (22%)  1.090                                                                   (0.138) The government party: anti-government grieving 4 (11%) 6 (67%) -1.322  (0.093) The political system and the government party: anti-system and anti-government grieving 1 (3%) 1 (11%) -0.234  (0.408)  The opposition party: grieving (pro-system) 2 (6%) 0 (0%) -78 Not attributed 3 (9%) 0 (0%) - Total negative descriptive claims 35 9 - Positive descriptive claims attributed to: SNP/SNP MPs: credit-claiming OR promises of future action  9 (20% of all positive claims) 15 (94%) -2.004  (0.022)  A large SNP vote: credit-claiming OR promises of future action 5 (11%) 0 (0%) -79 (Hypothetical) Scottish Government / Independence: positive case for self-government 6 (13%) 0 (0%) - Scotland’s natural resources (oil, forests, food production): positive case for self-government 1 (2%) 0 (0%) - No attribution 24 (53%) 1 (6%)  1.124 (0.131)  Total positive descriptive claims 45 16 -                                                  77 Source: Analysis of SNP election pamphlets and manifestos, 1983 and 2001. 78 Z statistics cannot be calculated in the last two categories there are no observations for 2001, meaning the variance is 0.  79 Z statistics cannot be calculated in the middle three categories as there are no observations for 2001, meaning the variance is 0.  119 only one party is named, this is coded as an attribution to the government party and opposition party respectively.  The key finding is that in 1983, the SNP attributed most of its negative descriptive claims to the overall political system. In 2001 the government of the day (Labour) was more likely to be blamed - the difference between the two campaigns’ anti-government claims is statistically significant at the 90% confidence level (though not the difference in the use of anti-system claims). The following examples show the difference between these two types of grieving. In the first, from 1983 when the Conservatives were in power, the SNP argues that oil revenue has been mis-used and frames this as a regional conflict between Scotland and the central government: “We warned that if London controlled Scottish oil, it would be frittered away and Scotland would lose one of its most exciting opportunities in history. How much has London squandered?” SNP, 1983, Leaflet 3, sentences 4-5 By contrast, in the second quote from 2001, when Labour was the incumbent government, the blame is attributed in partisan rather than geographic terms: “Or you can vote for more of the same tired excuses from Labour. Excuses for the fact that our pensioners have amongst the lowest living standards in Europe.” SNP, 2001, Leaflet 1, sentences 5-6  Furthermore, in 1983 the SNP attributed some problems to the Labour party alone, even though it had not been in power for four years, and had no legislative leverage in the 1979-1983 Parliament. In the quote below, the negative claim about Scotland's socio-economic problems is immediately followed by a claim that Scottish Labour MPs do not do a good job of representing Scotland. There is no mention of the (Conservative) government of the day at all:                                                                                                           120 “So when I see Scotland – the fifth largest oil producer in the world – suffering poverty, social deprivation, cuts in social service, and an appalling level of unemployment, I know that we could make a better job of running Scotland if we did it ourselves. Everybody knows that in the last Parliament 44 Labour MPs did nothing to defend Scotland.” SNP, 1983, Leaflet 1, sentences 14-15. Emphasis in the original.       Conversely, when Labour was in power in 2001, the SNP does not blame any problems on the Conservative opposition. This is so even though the Conservatives had been out of power for the same length of time as Labour in 1983.  These patterns of attribution fit with the overall argument that electoral competition drives separatists' rhetoric. When the SNP’s primary competitor is in government, its negative claims focus much more specifically on that government, as opposed to the overall system. Indeed, the separatists sometimes blame their electoral rival for Scotland's problems even when that rival is no longer in power.     So, we have established a difference between the extent and type of grieving in the two campaigns. However, the theory also predicted the use of credit-claiming in the 1983 campaign, when the SNP was competing with the official opposition party. The bottom half of Table 4.1 categorizes all positive descriptive claims (focusing on the present and the future) on the basis of who or what is given the credit.  In 1983, most positive claims were not attributed, but of those that were, the most common explanation given was that SNP MPs at Westminster had secured benefits for Scotland. Examples of such claims include:     “Everybody knows that sending SNP MPs to Parliament brings results for Scotland.” SNP, 1983, Leaflet No.1, sentence 17.   and  121 “In the last Parliament, with only 11 Scottish Nationalist MPs, Scotland got more ACTION and attention than we have ever had – BEFORE or SINCE! THINGS ONLY BEGIN TO HAPPEN FOR OUR COUNTRY WHEN THERE IS A LARGE NATIONALIST VOTE!” SNP, 1983, Leaflet No. 3, sentences 17-18. Emphasis in original.  The last example is particularly telling because it is actually claiming credit for SNP achievements in the 1974-79 Parliament (when it had 11 MPs) not the Parliament immediately preceding the 1983 election (when it had only 2 MPs). In other words, the SNP did have opportunities to gain concessions for Scotland when it was pivotal, in 1979, but it does not use this argument in its 1979 election campaign material. As will be shown in sub-section 4.3.2, the SNP’s 1979 campaign materials did not make any claims which attributed credit to its own MPs. Instead, it waited until the next electoral cycle to make this claim, when its electoral rival was no longer in power. This supports my argument that credit-claiming is less about separatist parties actually having the opportunity to gain concessions and more about who they are competing against. The SNP did not want to claim credit in 1979 when that would mean sharing it with a Labour government, but it did in 1983 when the incumbent government was not an electoral threat.                                                                                                                                        As Table 4.1 shows, in 2001, all but one of the positive claims were attributed to the SNP at Westminster. This is a significantly higher proportion than in 1983, which does not fit the hypothesis. The SNP is not pivotal in this election, and it is competing against the government, so it should not be credit-claiming. However, only one of the SNP’s positive claims in 2001 was about the present. In other words, these claims are not about benefits that SNP MPs did secure for Scotland during the 1997-2001 Parliament and under a majority Labour government. Rather, they are prospective claims about what the SNP will achieve in the future, as the example below shows:  122 “Protect Scotland from the worst failings of Westminster rule – the more SNP MPs, the greater chance we have of stopping the next fuel tax hike. As the SNP presence at Westminster increases so will the attention London pays to Scotland’s needs.” SNP, 2001, Leaflet No.2, sentence 30-31  This promise of future action does not disprove the hypothesis, as the SNP is being careful not to share any credit with the government of the day, its electoral rival. Rather, this statement shows that the SNP is appealing to different groups of voters simultaneously: to benefits-seeking voters it says that the SNP will demand attention to Scotland’s needs, to protest voters it remind them of an unpopular government policy (the fuel tax increase) and it also attacks the overall Westminster system in order to reinforce its separatist credentials.   In summary, then, the SNP’s electoral strategy shows that separatists parties can and do combine grieving and credit-claiming. In the 1983 campaign we see that grieving and credit-claiming are not mutually exclusive. The SNP continued to advocate for independence but also argued that while Scotland was still part of the UK, SNP MPs secured benefits for the region through their parliamentary work. Equally, in 2001 it combined a pro-independence message with the claim that its MPs can secure benefits in the future - if elected in sufficiently large numbers to the very legislature that the party rejects. The SNP doesn’t appear to be concerned that appealing to benefits-seeking voters or protest voters will alienate its core separatist support. It employs a mixture of pro-system and anti-system rhetoric, but the way in which it does so changes between elections. The party uses direct credit-claiming when its chief rival is in opposition. When that rival is in government, the Nationalists switch to anti-government grieving (likely to appeal to protest voters) and promises of future benefits (to appeal to benefits-seeking voters). This fits with the hypothesis that when self-government is off the agenda, it is electoral competition that drives separatists' rhetoric.  123 Given the long gap between these two cases, it is worth considering alternative explanations. Besides the change in government, there are two other salient differences between 1983 and 2001. First, the UK in 1983 was just coming out of a recession and unemployment was high: 12.7% in the month before the election (Government Statistical Service, 1996).  In 2001, the economy was doing well, with unemployment at 5% in the month before the election (Office for National Statistics, 2001).  However, if this factor was affecting SNP rhetoric, we would expect to see the opposite trend: more credit-claiming in 2001 when the economic picture was positive. Second, in 1983 the UK had a highly centralized government (Lijphart, 2012) while Scottish devolution had been established by 2001. This fits with the greater use of anti-system grieving in 1983: the separatist party attacks the political system most when its region has no form of self-government. However, this does not detract from the fact that the SNP replaces its anti-system grieving with anti-government grieving (as opposed to credit-claiming). If anything, the fact that the SNP attacks the government that granted devolution shows the importance the separatists attach to electoral rivalry. The SNP attacks a Labour government (its rival) more than it does a Conservative government – despite the better economic circumstances in 2001, and despite the Labour government’s constitutional concessions.  4.2.1.2 The IPP: 1906 v 1909-10 Prior to the 1906 and 1909-10 elections, the government of the day had a majority in the House of Commons, so the IPP was not pivotal. However, in 1906 the incumbents were the Conservatives, who ran Unionist candidates against the IPP in Irish constituencies.80  In 1909-10 the Liberals were in power, so the IPP’s electoral competitor was the Official Opposition.  In these two elections, the IPP was therefore in the same strategic position as the SNP in the 1983 and 2001 elections.                                                    80 Although Campbell Bannerman, the Liberal leader, was Prime Minister when the election was called, he had only just been appointed after Arthur Balfour’s Conservative government resigned. On appointment, Campbell-Bannerman immediately dissolved Parliament without any legislative activity. So, the governmental record of the previous 6 years was Conservative/Unionist, i.e. the electoral rivals of the IPP. 124 In 1906, the tone is overwhelmingly negative: 82% of the IPP’s descriptive claims point to problems in Ireland. In the 1909-10 manifesto, there are very few descriptive claims, but they are all employed to point out positive benefits.81  This is in line with the hypothesis, suggesting grieving when separatists compete against the government and credit-claiming when they compete against the opposition.  Table 4.2, below, shows who was blamed for the negative descriptive claims in the 1906 manifesto (there were no negative claims in the 1909-10 manifesto). There is an almost even split between anti-system grieving and anti-government grieving. Unlike the SNP, the IPP's anti-system grieving does not include attributions to both mainstream parties, nor do the Irish Home Rulers argue that no British party can be trusted. Instead, these anti-system claims are all focused on the institution of “Irish government” or “British government in Ireland”.   All the claims which blame the government of the day do so in partisan terms, but use the label “Unionist” as opposed to “Conservative”, and the tone is extremely negative, as in this example: “Coercion, decay, universal discontent – these words characterize the whole course of the Unionist government in Ireland.” IPP Manifesto 1906, sentence 7, The Times, January 1, 1906.  The intensity of this grieving is mirrored in the reasons given for voting IPP. The most common reason was to avoid another Conservative-Unionist government. This is mentioned five times in a manifesto which is only 24 sentences in length. It is not only recommended to the reader, but cast as his patriotic duty: “The first duty then of the Irish voters of Great Britain is to aid to the utmost of their power in the discomfiture of the great coalition [of Conservatives and Liberal Unionists] which has inflicted such immense injuries on their country.”  IPP Manifesto 1906, sentence 17; The Times, January 1, 1906.                                                  81 The difference between the distributions of claims in each election is statistically significant at the 95% level: the chi-square statistic is 10.5882 and the p value is 0.005. 125 Table 4.2 Descriptive claims and attributions made in the 1906 and 1909-10 IPP manifesto Negative claims attributed to: 1906 1909-10  The British political system  anti-system grieving 6 (43%) 0 The government of the day anti-government grieving 7 (50%) 0 No attribution  1 (7%) 0 Total negative claims 14 0 Positive claims attributed to:   IPP and Labour party – future actions 2 (100%) 0 IPP MPs – past actions 0 1 (33%) British Prime Minister & governing party 0 2 (67%) Total positive claims 2 3 Source: Analysis of IPP election manifestos, reported verbatim in The Times, 1906 and 1909-10   By contrast, in 1909-10, when the Liberals were in power, there are no attacks at all on the incumbent government’s record. When the party gives explicit reasons to vote IPP, there is no mention of beating electoral rivals or protesting against the incumbent government. So, the IPP has an even clearer partisan bias than the SNP. Although the Scottish Nationalists grieve against Labour (their rival) more than the Conservatives, there were some examples of grievances raised against both mainstream parties. The IPP focuses all of its grievances either on governmental institutions or the Unionists, its electoral rivals.   We therefore have evidence of grieving in 1906, when we expected it. The second prediction is that we should see credit-claiming in 1909-10, when the IPP's rival is out of power, but none in 1906 when the rival had been in office. In fact, both manifestos show evidence of credit-claiming, as positive claims are attributed to the efforts of the IPP’s Members at Westminster. However, the causal mechanism by which the IPP says it secured benefits for Irish voters differs.  126 In 1906, these claims are promises of future action, similar to the claims made by the SNP in 2001, when it too was competing against the incumbent government. The 1906 claims focus on Catholic education across the whole of the UK, as this affected Irish voters living outside of Ireland. The manifesto states what IPP politicians could achieve in the next Parliament, if their numbers were strong enough: “So far as the interests of the schools . . . we would urge upon the Irish National voters in Great Britain that the best means of protecting their schools is to strengthen the power and prestige of the Irish party.” IPP Manifesto 1906, sentence 18; The Times, January 1, 1906.   By contrast, when discussing the same issue in the 1909-10 manifesto, the IPP is no longer talking prospectively, but is drawing attention to its parliamentary activities on the topic in the previous parliament: “Need we say that on this important question the Irish party stand where they stood in the last Parliament, and that in giving unqualified support to that party the Irish in Great Britain are taking the most effective means to protect the Catholic schools.” IPP Manifesto 1909, sentence 14; The Times, December 17, 1909. Emphasis added.  This evidence points to a change in electoral strategy that matches that of the SNP. Both parties grieve when their electoral rival is in power. Both claim credit for actual parliamentary achievements (as opposed to promises for the future) when their rival is out of power. However, electoral strategy is not the only explanation for the IPP's shift in behaviour.  The IPP’s electoral rivals (the Conservative Unionists) were adamantly against Home Rule whereas the British Liberal party had recommitted itself to the policy during the 1909 election campaign. It is therefore difficult to disentangle whether the IPP was seeking to advance its first priority of self-government or whether it was focusing on election strategy. Attacking the Conservatives is consistent with both of these goals. The SNP's behaviour, however is more clear-cut: it attacks the Labour party more than the Conservatives. This is despite the 127 fact that Labour attempted Scottish devolution in 1979 and implemented it in 1999, whereas the Conservatives opposed Scottish self-government on both occasions. The SNP case therefore provides stronger evidence than the IPP case that separatist rhetoric can be driven by electoral competition – and that these concerns are not simply reducible to the pursuit of self-government.   4.2.2 When the separatist party is pivotal This sub-section examines the impact of electoral competition on rhetoric when separatists are pivotal: will the party always use anti-government grieving when its rival is in power, even if it has the opportunity to extract concessions? As noted in section 4.1, we do not have a clear-cut pair of elections which fits this scenario. To fill this gap, I look at the University of Aberdeen's collection of SNP campaign leaflets for Aberdeenshire constituencies in the 1979 election. Going into this election, there was a minority Labour government, so the SNP was pivotal. In that year Aberdeenshire had a range of different types of contests. There were two seats (Moray & Nairn and Aberdeenshire East) were SNP incumbents lost to the Conservatives. In both the previous election (October 1974) and the 1979 election, the top two parties in these seats were the SNP and the Conservatives, so the SNP is competing with the Official Opposition.82  There were also three seats where competition was between the UK-wide parties. If the theory from Chapter 2 is correct, we should see some credit-claiming in the seats where the SNP was competing with the Conservatives. In seats where Labour was a contender we would expect to see anti-government grieving.   Table 4.3, below, sets out who the SNP blamed for problems in Aberdeenshire, showing a clear difference between the Conservative v SNP seats and the seats where UK-wide parties were the main                                                   82 The categorization of constituency-level contests, in both the October 1974 and the 1979 elections is taken from C. Rallings and M. Thrasher, British Electoral Facts 1832-1999, 2000. 128 Table 4.3 Negative descriptive claims in SNP pamphlets, Aberdeen, 1979  Moray and Nairn Aberdeen-shire East Aberdeen-shire West Aberdeen North Aberdeen South SNP’s main competitor Opposition party Opposition party Opposition party Gvt party Gvt & Opp parties Negative state of affairs attributed to:  Conservatives/ Thatcher grieving; (pro-system) 17 (49%) 0 (n/a) 2 (9%) 0 (0%) 0 (0%) Labour/Government/Callaghan anti-government grieving 0 (0%) 0 (n/a) 1 (4%) 9 (56%) 4 (18%) Both parties anti-system grieving 4 (11%) 0 (n/a)  5 (22%) 1 (6%) 7 (32%) Westminster/London/UK anti-system grieving 0 (0%) 0 (n/a) 5 (22%) 2 (12.5%) 4 (18%) Other 10 (29%) 0 (n/a)  6 (26%) 1 (6%) 4 (18%) No attribution 4 (11%)         0 (n/a) 4 (17%) 3 (19%) 3 (14%) Total  35 0 23 16 22 Source: Analysis of SNP election pamphlets, 1979.  contenders. In Aberdeenshire East, held by the SNP, there were no negative descriptive claims at all, so this campaign appears to have not used grievances to appeal to the electorate.  In Moray and Nairn, held by the SNP's Winnie Ewing, nearly half of the negative claims were linked to either the Conservative Party or its leader, Margaret Thatcher. This was far more than for any other group or individual. This is despite the fact that the Conservatives were not in government and that Thatcher had (at that point) never been Prime Minister. Strikingly, none of the negative claims were attributed to the Labour government or its Prime Minister, James Callaghan.   129 Compare this to the constituencies where the SNP was not a contender. They all did make negative claims, but these are not focused on the Conservatives. Only two negative claims were attributed to the Conservatives in Aberdeenshire West and none in the remaining two constituencies. This suggests that the SNP did not blame problems on the Tories in all the constituencies where the latter were a contender – only in a constituency where the Tories were threatening an SNP incumbent.  By contrast, Labour was the target of most negative claims in Aberdeen North, a safe Labour seat, even though the SNP did not have a chance of winning that seat. As the quote below shows, the rhetoric used is aimed to undercut Labour's credentials for representing Scottish interests at Westminster: “In the last two years alone one million acres of Scotland have been sold to overseas buyers! With a record like this no wonder even one Labour MP was moved to call his colleagues ‘traitors to Scotland’” SNP, Aberdeen North, 1979, Leaflet 14, sentences 12-13  In sum, negative rhetoric was focused on the Conservatives when the SNP was directly competing with them, but anti-government grieving was used in a safe Labour seat. In seats where there was a close contest between UK-wide parties (Aberdeenshire West and Aberdeen South) the SNP spread the blame around, between both main parties and the political system itself. This suggests that at the constituency-level, the SNP adjusted its overall campaign to local electoral circumstances, particularly when it had seats at stake – giving further support to the argument that electoral competition drives rhetoric more than legislative status does.   Table 4.4, below, turns the attention to positive descriptive claims, setting out the groups and institutions that the SNP gave credit to. There is again a clear difference between the SNP's rhetoric in the SNP v Conservative seats and the contests between mainstream parties. In the former seats the SNP gives itself and in particular the incumbent MP credit. In the latter, benefits are either not attributed at   130 Table 4.4 Positive descriptive claims in SNP pamphlets, Aberdeen 1979  Moray and Nairn Aberdeen-shire East Aberdeen-shire West Aberdeen North Aberdeen South SNP’s main competitor Opposition party Opposition party Opposition party Gvt party Gvt & Opp parties Positive state of affairs attributed to:  Incumbent SNP MP credit-claiming OR promises of future action 2 (29%) 7 (87.5%) 0 (%) 0 (%) 0 (%) SNP MPs/party credit-claiming OR promises of future action 1 (14%) 1 (12.5%) 1 (11%) 1 (8%) 1 (6%) Scottish servicemen 2 (29%) 0 (0%)  0 (%) 0 (%) 0 (%) European Economic Community (EEC) 1 (14%) 0 (0%) 0 (%) 0 (%) 0 (%) Independent Parliament/Self-government positive case for self-government 0 (0%) 0 (0%)  6 (67%) 3 (25%) 8 (47%) Scotland  positive case for self-government 0 (0%) 0 (0%) 1 (11%) 0 0 No attribution 1 (14%)         0 (0%) 1 (11%) 8 (67%) 8 (47%) Total  7 8 983 12 17 Source: Analysis of SNP election pamphlets, 1979.  all or linked to some form of self-government. In other words, the positive claims in these constituencies were the positive case for separatism, not a recognition that Scotland could gain some benefits under existing arrangements.                                                  83 The Aberdeenshire West figures exclude a positive descriptive claim which was a quotation of local Conservative politicians. This claim praised the effects of the EEC's Common Agricultural Policy and was juxtaposed by the SNP's own negative claims about the impact of this policy. Quoting the Conservative claim was therefore intended to show that party as being out of touch, not to persuade the reader that Scotland was receiving benefits from Europe.  131 One example of credit-claiming in the Moray and Nairn leaflets concerned a campaign that Winnie Ewing led to improve pay and conditions for Scottish armed forces: “Mrs Ewing, in close liason [sic] with the Forces Wives' Association has identified herself with the battle for better remuneration and conditions for all Service Personnel and their families and as a result of this a big increase has been made. (She pointed out that the MOD [the UK Ministry of Defence] had underspent by £96.8 million last year.)” SNP, Moray and Nairn, 1979, Leaflet 29, sentences 2-3  What is interesting is that as well as claiming credit for an increase in forces wages, she is also attributing blame for the implicit claim that wages were too low before her intervention – by pointing to an underspend by the relevant government department. However, this is attributed to a bureaucratic body, the Ministry of Defence, rather than the party running the government, which was the Labour party. Not only does she abstain from blaming Labour, but later in the same leaflet she links her success in improving conditions to her co-operation with other parties: “Mrs Ewing pressed in Parliament for representatives of the Forces Wives' Association to be appointed to the Forces Pay Review Body, and for a Forces Ombudsman to be appointed. This received all Party support.” SNP, Moray and Nairn, 1979, Leaflet 29, sentences 18-19  This is in stark contrast with the previously quoted attacks on Labour MPs which were in nearby Aberdeen North's leaflets, in the same election (see above). The difference between the two constituencies is that Aberdeen North was a safe Labour seat whereas Labour was not competitive in Moray and Nairn. In a different issue area, that of inheritance tax, we see a similar pattern. Ewing identifies a problem, claims credit for her party's parliamentary activities to alleviate it, but avoids blaming the government of the day for the problem’s existence in the first place. Indeed here, she admits the SNP would like to have achieved more and blames the Conservative opposition (her electoral rivals) for that failure. This ignores the fact that if Labour had supported the SNP's proposals, 132 there would have also been enough votes in the Commons to pass the measure, and as the government of the day, Labour could have allocated more parliamentary time to the issue:  “The SNP recognized that Death Duties were a tax of the middle classes which the very rich could always find means to avoid. . . .Accordingly we pressed for much higher thresholds for Capital Transfer Tax and again the success was limited by the unwillingness of the Conservative Party to go as far as we would have liked.” SNP, Moray and Nairn, 1979, Leaflet 27, sentences 17 and 19. Emphasis added.  In the other SNP v Conservative seat, the leaflets also contained credit-claiming. These claims focus on the SNP's ability to co-operate with other parties in order to obtain benefits for Scotland, covering issues from fishing zones to gas plants. They also quote a local newspaper columnist who summed up the claim that the SNP acted as Scotland's regional champion at Westminster during the 1974-79 Parliament, and that the party was able to extract benefits from the government: “ 'The presence of SNP MPs at Westminster during the period of the last Parliament did more than any other factor to galvanise Government Ministers into taking account of the Scots.' John Deans (Press & Journal 10 April 1979)”                                     quoted in SNP, Aberdeenshire East, 1979,  Leaflet 20, sentence 22.  What we see in both of these constituencies, therefore, is credit-claiming designed to appeal to benefits-seeking voters. This rhetoric has an implicitly pro-system message, as it gives examples of how Scottish representatives can gain benefits for their local communities by participating in the central legislature and co-operating with other parties. We do not see such arguments in the other constituencies, where the SNP was not under pressure from the Official Opposition – there we see both anti-system grieving, which appeals to the core separatist vote and anti-government grieving which can attract protest voters. These conclusions fit with what we would expect if electoral competition was the primary determinant of SNP rhetoric.  133 One counter-argument to this conclusion is that in Moray and Nairn and Aberdeenshire East there were incumbent SNP MPs. So, credit-claiming may have simply been about defending their personal record rather than a reaction to the fact they were competing with the opposition party (and not the government). However, in Moray and Nairn this credit-claiming was coupled with a focus on blaming the Conservatives for Scotland's problems, as opposed to the critique of the Labour government and the Westminster system that we see elsewhere. This combination of credit claiming and anti-Tory arguments suggests that the SNP was not simply trying to defend its incumbent MPs: rather, it was cognizant of which party was a threat to those MPs. As we will see in sub-section 4.3.2, the SNP's overall campaign in 1979 was premised on anti-government and anti-system grieving. This fits with my theory, because its primary competitor in Scotland as a whole was the government.   However, these constituency level data show that when individual candidates were competing with the opposition, they used very little anti-government grieving and instead incorporated a localized credit-claiming strategy. This mirrors the Scotland-wide campaign in 1983, when the SNP's primary rival was in opposition (see sub-section 4.2.1, above).   4.3 Testing the impact of legislative status Having considered the impact of electoral competition, this section moves on to the second independent variable, legislative status, and the cases set out in Figure 4.2. Here, I compare elections when the separatist party had been pivotal in the previous parliament to ones where it had not. The research question is whether such an opportunity to gain concessions makes a difference: are separatists more likely to credit-claim in such circumstances? The theory predicts that this will not make a difference. Rather, rhetoric will still be primarily determined by the identity of the separatists' electoral competitor. So, they will grieve when competing with the government of the day and credit-claim when competing with the opposition, regardless of the separatist party's status in the previous parliament. 134 4.3.1 When separatists are competing with an opposition party 4.3.1.1 The IPP 1909-10 v December 1910 In its heyday the IPP had a larger contingent of MPs than the SNP has ever had (even including the 2015 UK elections results). This meant that in the late nineteenth and early twentieth centuries Britain had a three-party system. So, examining the IPP allows us to test whether pivotal status had an impact on electoral behaviour in a multi-party system, which makes minority government more likely.  The 1909-10 election followed a period of majority government and is therefore a useful comparator to the election held a year later in December 1910, when the IPP was pivotal. In both of these elections, the incumbent government was Liberal, meaning that the IPP's primary competitor, the Conservative-Unionists, were in opposition.   Table 4.5, below, sets out how descriptive claims were used in the two manifestos. The 1909-10 manifesto fits the pattern we would expect if the party was credit-claiming: all descriptive statements about the state of affairs in Ireland are positive, both when talking about the present and the future. The December 1910 manifesto does not fit this pattern as it splits its descriptive claims evenly between positive and negative claims.  All the positive claims are actually about the future and hence less concrete than the claims made in the 1909-10 manifesto. The differences between the two years distributions are significant at the 10% level (the chi-square statistic is 6.9667 and the p value is 0.073).84                                                    84 I use the 10% level for hypothesis testing for the IPP manifestos because they are short, meaning that there are very few observations (i.e. descriptive claims) to work with. Substantively speaking, there is a clear difference in the tone of the two manifestos. 135  Table 4.5 Descriptive claims made in IPP electoral campaign materials Negative claims attributed to: 1909-10 Dec 1910  The House of Lords  anti-system grieving 0 3 (75%) The government of the day anti-government grieving 0 0 Unionism grieving against the opposition party 0 1 (25%) Total negative claims 0 4 Positive claims attributed to:   Working class voters 0 2 (50%) IPP MPs  credit-claiming 1 (33%) 2 (50%) British Prime Minister & governing party credit-claiming 2 (67%) 0 Total positive claims 3 4 Source: Analysis of IPP election manifestos, reported verbatim in The Times, 1909-10 and Dec 1910  This is despite the fact that the IPP is still competing against the opposition. The party should therefore have an electoral incentive to credit-claim, in order to appeal to benefits-seeking voters. This group of the electorate is less likely to vote for the Conservatives when the latter are not in power, so you would  expect them to be a tempting target for the IPP. Furthermore, the IPP has more opportunity to credit-claim in the later election than it does in 1909-10, as its pivotal status makes claims that it secured benefits for Ireland more credible.   Looking in more detail at the content of these descriptive claims will help to ascertain why the December 1910 campaign is negative despite the IPP’s apparent advantages. The negative claims are not attributed to the incumbent Liberal government, rather, 3 are blamed on a political institution. This is the House of Lords, which as the quote below explains to voters, was blocking Irish Home Rule: 136 “For 20 years the House of Lords has stood between the Irish people and full self-government, and now, exasperated by oft-repeated injuries and insults, the long-suffering democracy of Great Britain has risen in revolt against that ancient fortress of privilege and monopoly.” IPP Manifesto December 1910, sentence 5, The Times, November 30, 1910.  So, this grieving is anti-system and underlines the fact that even though Home Rule was not on the agenda in 1910, it still dominated the IPP's rhetoric and priorities. What we do not see are attacks on the government of the day, which fits with the theory as the Liberals were not an electoral threat to the IPP. Moreover, the fact that grieving (of any kind) occurs in December 1910 supports the prediction that pivotal status does not make a difference. If it did, we would expect to see more credit-claiming and less grieving in comparison to the previous year's campaign, and in fact we see the opposite.  The next question to ask is whether the positive descriptive claims noted in Table 4.4 constitute credit-claiming, as my theory predicts. In both years, they are attributed to the efforts of the IPP itself to represent and “save” the Irish. For example: “The Irish Party stands today where is has always stood. . . It [the IPP] is not responsible to any one man or to any section of men, but to the whole Irish people whose creation it is, and whose salvation it has proved itself to be.” IPP Manifesto December 1910, sentences 25 and 27, The Times, November 30, 1910.  However, with the exception of the schools issue (see sub-section 4.2.1 above), the credit claiming that occurs in the 1909-10 and December 1910 manifestos is all tied to Home Rule. Exit from the Westminster system was not an immediate prospect in either 1909 or 1910, due to the House of Lords' veto.  Nevertheless, the IPP structures its appeals for votes around its ability to get movement on the issue: “The one duty of this country [Ireland] is to return to Parliament a pledge-bound Irish Party, stronger and more united than ever, so that every other consideration must be sacrificed.” IPP Manifesto December 1910, sentence 17, The Times, November 30, 1910.  137 The pledge referred to here is the written agreement that all IPP candidates signed to support Irish Home Rule. Furthermore, the IPP does not simply rely on its own impact in Parliament, but instead builds electoral alliances, by encouraging Irish voters in the rest of the UK to support candidates sympathetic to its cause: “The only question we advise to be put to candidates is: “Do you accept the declaration of policy [linking Irish Home Rule to House of Lords reform] made by the Prime Minister at the Albert Hall on December 10?” IPP Manifesto 1909-10, sentence 7, The Times, December 17, 1909.  Comparing these elections therefore leads to similar conclusions to the comparison of the 1906 and 1909 elections (in sub-section 4.2.1, above). The IPP manifestos do fit the predictions of Chapter 2. First, the party credit claims when its rival is in opposition, and it does not grieve against the government of the day. Second, pivotal status does not lead to an increase in this use of credit claiming. Electoral competition therefore appears to have an effect whereas legislative status does not. However, this does not mean that the party is putting short-term electoral advantage over long-term pursuit of its goals. For the IPP in this period there is no trade-off between the two.   4.3.2 When separatists are competing with a government party 4.3.2.1 The SNP: 1979 v 2001 Comparing the SNP’s behaviour in the 1979 and 2001 elections isolates the impact of pivotal status, as there was a minority Labour government in the run-up to the former election and a majority Labour government in the run-up to the latter. As Labour is the SNP's primary competitor for Scottish seats, we would expect to see the SNP grieving in both elections.  However, if my theory is wrong and legislative status is important, we will see the SNP grieve more when it was not pivotal (in 2001). This is because it cannot credibly claim to be responsible for anything that is going well.   138 Table 4.6, below, shows that negative claims were used in both campaigns, but there was a shift from anti-system grieving in 1979 to anti-government grieving in 2001. In 1979, the SNP blamed nearly all of Scotland’s woes on “London, “Westminster”, “England” or on all the nationwide parties, suggesting that being subject to a UK-wide government was the root cause. This logic is elaborated in one pamphlet on local housing, which first details a problem (i.e. making a negative descriptive claim):   “Vast new schemes (Westhill, Elrick, Dyce, Kemnay, Bridge of Don, Kintore) are being built without adequate amenities such as shops, schools, medical services, halls, churches, post offices, banks and parks which Scots people have a right to expect.”  SNP, 1979, Leaflet No. 21, sentence 2. Then in the next sentence, the SNP attributes this to economic problems in the rest of the UK: “Meanwhile the Oil Wealth, (worth £5 million per day) is flowing South to prop up the bankrupt government in Westminster.” SNP, 1979, Leaflet No. 21, sentence 2. Emphasis in original.  Note that the government is described using institutional and regional terms (“Westminster”; “South”) rather than partisan labels. As Table 4.6 shows, the 1979 leaflets do not attribute any of the negative claims to the government of the day specifically. When party labels are mentioned, both main parties are blamed for the problem at hand, as in this leaflet on Scotland’s universities, starting with a negative empirical claim:  “It had become very difficult to attract the best graduates and many of the most able staff had resigned.” SNP, 1979, Leaflet No. 22, sentence 8. which is attributed as follows: “While the present mess was chiefly due to the irresponsibility of the Labour Party, there could be little hope for improvement under any Tory regime committed to draconian public expenditure cuts.” SNP, 1979, Leaflet No. 22, sentence 9.    139 Table 4.6 Negative descriptive claims and attributions made in SNP electoral materials Attributed to: 1979 Pivotal; Government competitor 2001 Not pivotal; Government competitor Z statistic for the 1979 v 2001 difference British political system: anti-system grieving 40 (78%) 5 (42%) 0.829 (0.204) Government of the day (Labour party): anti-government grieving 0 (0%) 7 (58%) -85 English academics at Scottish universities Grieving 5 (10%)  0 (0%)  - No attribution 6 (12%) 0 (0%) - Total  51 12 - Source: Analysis of SNP election pamphlets and manifestos, 1979 and 2001.  These attribution statements suggest that a change of government will not alleviate problems. They are then linked to the SNP’s overall solution: an independent Scotland. This argument can be seen in both the SNP’s policy prescriptions and its appeals to individual voters. In 1979, there were 17 explicit reasons given for voting SNP; 11 of these were to bring various benefits to Scotland which could only be achieved through independence. Leaflets which set out the SNP’s policy on land ownership and farming provide an illustration of the structure of this argument. They detail the party’s specific proposals in these issue areas, before asking the reader to “support this SNP policy with [local candidate name]” (SNP 1979, Leaflet 23 sentence 23 and Leaflet 24, sentence 19). Both leaflets have a separate text box referring to the referendum on Scottish devolution which occurred (and failed) two months before the 1979 election:                                                  85 Z statistics cannot be computed for the latter three categories as there are no observations for either 1979 or 2001 in each category.  140 “Referendum Spot – Under present proposals farming policy will not be devolved and therefore is likely to be dominated even more by English needs than is now the case. Only when we are independent will we be able to have a farming policy suited to our situation and our needs.” SNP, 1979, Leaflet No. 23, sentences 20-21. and “Referendum Spot - Land ownership will remain under the control of Westminster. Policies detailed above will only be possible in an independent Scotland.”  SNP, 1979, Leaflet No. 24, sentences 17-18. Each leaflet thus poses substantive policy proposals as a reason to vote SNP, before linking those same proposals to the need for independence. In the process the party rejects even the part measure of a devolved Scottish Parliament which had been the subject of the 1979 referendum.  However, by 2001, there had been a shift towards anti-government grieving. Independence no longer formed a substantial part of the case for why a voter should support the SNP. Between the 2001 manifesto and leaflets, there were 92 separate reasons explaining why the reader should vote SNP, but only 10% of these were linked to support for self-government. Furthermore, the party no longer attributed all of Scotland’s problems to being part of the UK.  As Table 4.6 shows, in the 2001 election material the blame was split between the constitutional framework and the incumbent Labour government. An example of the latter is:  “Or you can vote for more of the same tired excuses from Labour. Excuses for the fact that our pensioners have amongst the lowest living standards in Europe.” SNP, 2001, Leaflet 1, sentences 5 and 6.  These results are in line with the prediction that a separatist party will grieve when its primary electoral competitor is in government. It also shows that pivotal status does not have an impact. If anything, when the SNP had been pivotal in 1979, its electoral rhetoric was more anti-system than in 2001. If legislative status was driving rhetoric, we would expect this pattern to be reversed: for the SNP to attack the political system more when it had no legislative leverage, and less when it did. Instead, by 141 2001, there is far more focus on blaming the SNP’s electoral rival for Scotland’s problems. However, the gap between anti-system attributions in the two years is not statistically significant, so we should not place much emphasis on this shift in the type of grieving between the two campaigns. What we can see clearly is that in both years there is a focus on grieving: as we would expect if electoral competition was the key factor affecting separatist campaigns.   4.3.2.2 The BQ: 2004 v 2005-06 and 2008 In each of the 2004, 2005-06 and 2008 campaigns, the Bloc's primary rival appeared to be the incumbent government.86  My theory therefore predicts that the BQ will use grievance-based rhetoric in all three elections. If, though, pivotal status is more important than electoral competitor, we should see the Bloc’s behaviour change between 2004 (when it was not pivotal) and the later elections when it was.   Table 4.7, below, shows who the BQ blames for Québec’s problems. There is an interesting pattern, suggesting that pivotal status does have some impact on the rhetoric of manifestos. Anti-government grieving was significantly more common in the elections when the BQ was pivotal, whereas anti-system grieving is more prevalent in 2004, when the Bloc was not pivotal.  For example, in 2004, the BQ blames the political system as a whole for regional inequality within Québec: “However, in recent years, a two-tier Québec has been emerging. This divides Québec, and Ottawa is largely responsible for this situation having abandoned some regions to their fate.” BQ Manifesto, 2004, p. 3, sentences 45-46.  By contrast, the Bloc blamed problems in a range of issue areas on the incumbent government in both 2006 and 2008. It did this even though the actual government had changed (from the Liberal party to                                                   86 This categorization is based on opinion polls from the six months prior to each campaign (see Appendix 1). The incumbent government was the Liberal Party in 2004 and 2005-06 and the Conservative Party in 2008. 142 Table 4.7 Negative descriptive claims and attributions in BQ electoral campaign materials  2004 Not pivotal; Government competitor 2005-06 Pivotal; Government competitor 2008 Pivotal; Government competitor  Z-statistic for 2004 v 200687  (p value) % of all negative claims attributed to the Canadian political system: anti-system grieving 17 (47%) 1 (4%) 0 (0%) 1.029 (0.152) % of all negative claims attributed to the government of the day: anti-government grieving 9 (25%) 22 (96%) 6 (100%) -1.861 (0.031) Total of negative claims 36 23 6 - Source: Analysis of BQ election manifestos, 2004, 2005-06, 2008.  the Conservative party). This suggests that the Bloc's rhetoric was not a reaction to the government's actual ideological position, but to the electoral threat posed.   One possible explanation for the switch from anti-system grieving to anti-government grieving is that it is easier to argue that a political system is not meeting your region’s needs when your own party is not pivotal: your MPs do not have a realistic chance to do anything about it. In the 2006-8 period, the Bloc                                                  87 Z statistics cannot be calculated for the differences between 2008 and the other campaigns, because there are no observations in the anti-system category and 100% of the observations in the anti-government category, meaning that the variance in both cases is 0. 143 did have the opportunity to extract concessions. Nevertheless, it had an electoral incentive to avoid such deals with the government. In such circumstances, therefore, if a party is going to raise grievances it also has to explain to its voters why it has not used its parliamentary leverage to address those problems. One way to do this is to depict the government of the day as negatively as possible, in order to explain why the separatist party is unable to broker deals with it. We see evidence of this in both 2006 and 2008. In the former election, the BQ actually outlines a potential deal offered by the minority Liberal government to ward of a no confidence vote and compares it to a corruption scandal that had cost the Liberals support in Québec: “Demonstrating the same arrogance that led to the [sponsorship] scandal, the Liberal government introduced a budget statement with which it once again tried to buy the vote of Québecers by a process akin to a little petty blackmail.” BQ Manifesto, 2006, p.3, sentence 20.  Similarly in 2008, the BQ depicts the Conservative government as fundamentally opposed to Québec’s interests, thus explaining why it was unable to co-operate with the new government: “The Harper government refused to take action to protect the French language and has made some disastrous decisions for our economy, our regions and Québec culture.”  BQ Manifesto, 2008, p.3, sentence 19.  However, as shown in Table 4.6 above, the SNP shows the opposite pattern. So although this explanation of the Bloc’s rhetoric makes intuitive sense, there is not strong evidence to suggest that separatist parties in general use anti-government grieving when pivotal and competing with the government and switch to anti-system grieving when not pivotal.   What we do have is evidence that the Bloc persisted with a grievance strategy of one kind or another throughout these elections, despite its pivotal status in the latter two. A final check on this conclusion is to examine the rare positive claims that were made about Québec, to see if any of these suggest credit-144 claiming. Table 4.8, below, examines all positive claims made by the Bloc in these elections. It shows that in all three years, most positive claims were attributed to the BQ itself, which would fit with credit-claiming.88 However, with the exception of two claims made in 2004, these claims are all promises of future action, of the type used by the Scottish National Party in 2001 (see sub-section 4.2.1). This is when the party sets out positive claims about the future, posited on things it will do if it has more MPs in the future. It is not credit-claiming for things it had actually achieved in the current political system.  Such claims do not share credit with the government of the day and thus do not detract from anti-government grieving. In fact, this example from 2008 shows how such a claim can be combined with an attack on the government of the day: “If we are to impose respect for Québec in Ottawa, we cannot let Stephen Harper’s Conservatives win a majority. . . A strong contingent of Bloc Québécois MPs in Ottawa is the best possible outcome of this election for Québec.” BQ Manifesto 2008, p.3, sentences 29-31.  There are only 2 sentences across three manifestos which meet the requirements for credit claiming: a causal claim that the separatists’ actions have secured benefits for its home region. These were in the 2004 manifesto: “During the 2000 election campaign, the Bloc put a lot of pressure on Ottawa regarding the adoption of an anti-gang law to put behind bars the biker gangs plaguing Québec. After a long struggle, such a law was passed and dozens of these bikers were sentenced to prison.” BQ Manifesto 2004, p.4, sentences 35-36.                                                    88 The differences between each of the three years were not statistically significant. The z statistic for 2004 (non-pivotal) v 2006 (pivotal) was -0.347 (p value 0.729 on a 2-tailed test); for 2004 (non-pivotal) v 2008 (pivotal) it was 0.000 (p value 1.000 on a 2-tailed test). The z statistic for the difference between the two pivotal campaigns, 2006 and 2008 was 0.331 (p value 0.741 on a 2-tailed test).  145 Table 4.8 Positive descriptive claims and attributions made in BQ electoral materials Attributed to: 2004 Not pivotal; Government competitor 2005-06 Pivotal; Government competitor 2008 Pivotal; Government competitor BQ/BQ MPs in Ottawa credit-claiming OR promises of future action 9 (60%) 3 (75%) 3 (60%) Québec sovereignty positive case for self-government  1 (7%) 189 (25%) 0 (0%) Québécois human resources (businesses, people, culture) positive case for self-government 2 (13%) 0 (0%) 0 (0%) Stopping a Conservative majority government anti-government grieving 0 (0%) 0 (0%) 1 (20%) Other 3 (20%) 0 (0%) 0 (0%) No attribution 0 (0%) 0 (0%) 1 (20%) Total  15 4 5 Source: Analysis of BQ election manifestos, 2004, 2005-06, 2008.  Here, the Bloc was not pivotal, either at the time of its actions on gang violence, or at the election when it sought credit – this provides further evidence that credit claiming is not dependent on pivotal status. The lack of its use when the Bloc was pivotal supports the conclusion that when separatists are competing with the government party, campaigns will focus on grieving, not credit-claiming.                                                   89 The z statistic for the difference between 2004 and 2006 was -0.391, p value 0.695 (2-tailed test). Z statistics are not calculated for 2008 as there are no observations in this category, nor for the other categories as they all have 0 observations for two of the three years. 146 4.4 Conclusion The findings of this chapter are summarized in Tables 4.9 and 4.10, below. This sets out the four scenarios examined in this chapter, linking these back to Table 2.2., in the theory chapter. The Table compares Chapter 2’s predictions in each of these scenarios to the evidence presented here. For each election, the party’s behaviour matches the predicted behaviour. This is most arguable for the SNP in 1983 which uses both credit-claiming and grieving. However, the party shows a marked drop in its use in grieving and an increase in the use of credit-claiming when compared to the 1979 election: both of these changes are in the hypothesized direction. This case does suggest that Chapter 2 draws too sharp a distinction between these two strategies. Rather than being mutually exclusive, parties do manage to craft appeals which could appeal to core separatists at the same time as floating voters.   The key point is that separatists actually do acknowledge that their region has received concrete benefits under the current constitutional arrangements. In particular, we see the SNP do this as a party in 1983 and in the campaigns of some individual MPs in 1979. This runs contrary to the predictions of the anti-system party literature, as it shows that separatists will, under certain circumstances use rhetoric which is actually pro-system. In line with my theory, this electoral behaviour varies with the status of the separatists' key competitor. The only instances of credit-claiming based on actual achievements (as opposed to promises for the future) occur when separatists are competing with an opposition party.  Pivotal status does not make credit-claiming more likely. These findings therefore suggest that when self-government is off the immediate agenda, separatists do take account of the position of other parties – however it is electoral competition that has the greatest impact, not legislative interaction.    147 Table 4.9 Predictions and conclusions: testing the impact of electoral competitor   Separatists are competing with the Government Separatists are competing with the Opposition Separatists are not pivotal in the legislature   SNP, 2001 IPP 1906            Predicted behaviour: Grieving  Findings: SNP: Grieving IPP: Grieving    SNP, 1983 IPP 1909-10  Predicted behaviour: Credit-claiming  Findings: SNP: Grieving & credit-claiming IPP: Credit-claiming  Separatists are pivotal in the legislature    SNP 1979: Aberdeen North constituency  Predicted behaviour: Grieving  Findings: SNP: Grieving & making case for separation       SNP 1979: Moray and Nairn and       Aberdeenshire East constituencies  Predicted behaviour: Credit-claiming  Findings: SNP: Credit-claiming           148 Table 4.10 Predictions and conclusions: testing the impact of legislative status   Separatists are not pivotal in the legislature Separatists are pivotal in the legislature Separatists are competing with the Opposition   IPP 1909-10            Predicted behaviour: Credit-claiming  Findings: IPP: Credit-claiming   IPP Dec 1910  Predicted behaviour: Credit-claiming  Findings: IPP: Credit-claiming Separatists are competing with the Government   SNP 2001 BQ 2004   Predicted behaviour: Grieving  Findings: SNP: Grieving BQ: Grieving; one example of credit-claiming    SNP 1979 BQ 2006 & 2008  Predicted behaviour: Grieving  Findings: SNP: Grieving BQ: Grieving            149 Chapter 5: When do separatists participate in the central parliament? This chapter focuses on separatists’ behaviour once elected: do they participate in an institution which they argue should not be making decisions about their region? Young and Bélanger’s survey of separatist parties in Western democracies finds a spectrum of behaviour, from outright boycott of the central parliament to participation in the arliament and the government of the day. Meanwhile, Blidook and Soroka et al’s work finds that mainstream Canadian politicians use Private Members’ Bills and parliamentary questions to raise their constituents’ concerns. The hypotheses set out in Chapter 2 argue that separatist parties adjust their legislative goals and their strategies according to two factors: whether self-government is on the agenda and whether they are pivotal in the legislature. If this is true, there will not only be variation between separatist parties. We would also expect the same separatist party’s MPs to change their legislative behaviour over time, acting more like mainstream politicians when self-government is off the agenda.           The first hypothesis states that separatist parties are less likely to participate in legislative activities when two factors are present: self-government is on the agenda and the separatist party is not pivotal. This is because engaging with the central legislature risks diluting the separatists’ message that their region should be autonomous. This risk cannot be offset because the separatists have no leverage to extract concessions – so I predict that they will disengage from central institutions. However, when self-government is on the agenda and the party is pivotal it has the opportunity to extract concessions from other parties and will use these to work towards its primary goal. The second hypothesis therefore predicts that in this scenario, separatist MPs will engage in parliamentary activities related to proposals for self-government.   150 Conversely, when self-government is off the agenda, the party expects to be tied to the existing political system for at least the medium term. We would therefore expect participation in the central legislature, but the separatist party’s focus will differ depending on whether it has any opportunities to seek concessions. The third hypothesis states that when the party is pivotal it will participate in discussions of internal institutional reform, in order to strengthen the position of small parties in the legislature. This is to ensure that the party will have the opportunity to seek goods and services for its region, which it can use to shore up electoral support. Finally, when the party is not pivotal it has few chances to gain such concessions, so will not bargain. However, the party’s MPs can still use the parliament as an authoritative platform from which to make the case for self-government and build its profile with the electorate. In this final scenario I therefore expect to see separatist Members use their legislative time to publicize their region’s grievances.   This chapter tests these predictions by comparing separatist parties’ legislative behaviour over time. I begin by setting out the parliamentary sessions which will be used to test these scenarios and explaining the use of pro-system parties as comparators to the separatist parties. These time periods were chosen to isolate the effect of the two independent variables: the separatists’ status in the legislature and whether self-government was on the agenda or not. The second section operationalizes the dependent variable, setting out the behaviour I should see if the hypotheses from Chapter 2 are correct. In the first of the results sections I examine the Bloc Québécois (BQ), during the first Parliament of the Chrétien Liberal government, which ran from 1993 to 1997. This allows us to compare the BQ’s behaviour before and after the 1995 referendum on Québec sovereignty.  The Bloc was not pivotal, so this section focuses on testing the first hypothesis on disengagement. The second results section focuses on the actions of the Scottish National Party (SNP) during the Callaghan Labour government of 1976-79. This period includes a failed attempt to devolve some legislative powers to a 151 Scottish Assembly. The Callaghan era therefore gives us another opportunity to study how separatists react to self-government falling off the agenda. However, unlike the BQ, the SNP was pivotal in this period, allowing us to test the second and third hypotheses on how separatist parties use legislative leverage and whether this changes with the prospect of self-government. Finally I return to the Bloc during the Martin Liberal government, from 2003 to 2005. Sovereignty was not on the agenda during this era, but it includes periods of both majority and minority government. This provides another test of hypothesis 3, on separatist involvement in institutional reform. However, it differs from the SNP comparison which holds pivotal status constant, asking if separatists engage more in the central legislature when self-government falls off the agenda. The Bloc comparison holds the prospect of self-government constant – so it asks whether separatists participate more when they become pivotal.  5.1 Case selection In order to capture temporal changes in a separatist party’s behaviour, I selected paired time periods from each party’s parliamentary history. I use four criteria to choose these cases. First, I want to isolate each of the independent variables, so I select pairs of time periods where one is held constant but the other changes.90 Second, the two time periods should be consecutive, so as to minimize variation in unobservable factors which might also affect a party’s behaviour (e.g. changing attitudes in the electorate at large, developments in communications technology, leadership etc.). Third, the governing party and Prime Minister should not change between these two time periods, so that the government’s own ideology and legislative agenda can be ruled out as a factor in the separatist party’s behaviour.   A final key point concerning selection is the presence of other opposition parties which can be used as comparators to the separatist party. Legislators do not have an unrestricted choice as to which debates                                                  90 See Chapter 3, section 1 for an explanation of how I define “a reasonable prospect” of self-government. 152 they will speak in, when they will introduce motions and when they will ask questions. The government’s control of the legislative agenda (Cox, 2005) means that there is a selection effect as to how often MPs speak, on which occasions and on which issues – this is particularly so for members of small opposition parties. Additionally, the largest opposition party may have very different incentives from the smaller opposition parties on the subject of institutional reform. The official opposition party sees itself as a contender for government, so will not want to agree to parliamentary reform which may hinder its own agenda when next in government (Horn, 1995). For these reasons, I compare the legislative activities of the separatist parties to those of small, pro-system parties and new parties of a similar size in the same parliament.  I define “small parties” as those that are not plausible contenders to lead the next government: in other words, any parties other than the party leading the current government or the official opposition.  In the event of a hung parliament, small parties might join a government as a junior coalition partner, but the head of government will not come from their ranks. In parliaments where the separatist party has won seats for the first time, I use comparator parties which are themselves new-comers. This means that the level of experience and institutional knowledge of the MPs is held constant.   These comparator cases give a baseline of expected behaviour from a party which does not want to leave the political system, but which faces a similar level of procedural disadvantage, and the same government agenda, as the separatist party. This allows us to test whether the separatists’ shorter time horizons have any impact on their behaviour. For the SNP’s comparators, I use the various Unionist parties in Northern Ireland, as well as the Liberal party. The latter was a small statewide party suffering from a co-ordination failure among voters at the centre of the left-right spectrum. Unlike the SNP, these comparator parties all supported the territorial status quo in the United Kingdom.   153 For the BQ’s comparators, I use the New Democrat Party (NDP) and the Reform Party. During the periods under examination, the NDP’s size was limited by co-ordination failure among centre-left voters, many of whom voted for the centrist Liberal party outside Québec and for the BQ itself inside Québec. The NDP was the fourth largest party in the Canadian Parliament, so, like the BQ, it was not a plausible contender for government. However, in the first two time periods of 1993-5 and 1995-7, the BQ was actually the second largest party in Parliament. It therefore had a much larger caucus and, as the Official Opposition, it had procedural advantages denied to the NDP. To account for these factors, I use the Reform Party as a second comparator: it was of similar size to the Bloc and was also a newcomer to the Canadian Parliament.   Figures 5.191 and 5.292 set out the time periods chosen using these selection criteria.93  Figure 5.1 shows that these pairs isolate the impact of a change in the prospects for self-government, both when the separatists are not pivotal (the BQ between 1993 and 1997) and when they are (the SNP between 1977 and 1979). Conversely, these cases also isolate the impact of a change in pivotal status (the BQ from 2003 to 2005): see Figure 5.2.                                                       91 Sources: derived from data from the Parliament of Canada, 2013 and Craig, 1989. 92 Source: derived from data from the Parliament of Canada, 2013. 93 The time periods which meet the four selection criteria are all from either the SNP or the BQ’s parliamentary history. In the IPP’s case, there is no other small party in the British Parliament until the 1900 election, at which the left-wing Labour party won seats for the first time. The IPP’s parliamentary activities during the 19th century are therefore excluded from analysis due to the lack of a comparator. During the party’s 20th century history, there is only one point at which one independent variable changes while the other stays constant. This is in 1915, when the IPP’s legislative status changes from pivotal to non-pivotal. However, this change was caused by a change in the composition of government (from a Liberal minority to a Liberal/Conservative coalition), thus violating the 3rd selection criterion.   154   Self-government on agenda Self-government off agenda Separatists not pivotal in legislature   BQ (18% of MPs)  from Federal election  (25 Oct 1993) to  No vote in sovereignty referendum (30 Oct, 1995)  Comparator parties: Reform Party  18% of MPs, regional party  New Democratic Party 3% of MPs, statewide party  Predicted behaviour: Disengagement  (Hypothesis 1)   BQ (18% of MPs)  from No vote in sovereignty referendum  (31 Oct 1995) to Federal election (2 June 1997)  Comparator parties: Reform Party  18% of MPs, regional party  New Democratic Party 3% of MPs, statewide party  Predicted behaviour: Using Parliament as a platform for grievances (Hypothesis 1) Separatists pivotal in legislature  SNP (1.7% of MPs)  from Government’s loss of majority  (24 Feb 1977) to  Beginning of Lib-Lab pact (23 Mar 1977).  Comparator parties: Liberal Party 2% of MPs & statewide party  Ulster Unionist Party 0.9% of MPs & regional, non-separatist party  Predicted behaviour: Bargaining for self-government  (Hypothesis 2)  SNP  (1.7% of MPs)  from No vote in Scottish devolution referendum (1 Mar 1979) to  Fall of Government (28 Mar 1979)  Comparator parties: Liberal Party 2% of MPs & statewide party  Ulster Unionist Party 1.1% of MPs & regional, non-separatist party  Predicted behaviour: Bargaining for parliamentary reform and regional benefits (Hypothesis 3) Figure 5.1 Paired cases, varying prospects for self-government; holding pivotal status constant  155   Separatists not pivotal in legislature Separatists pivotal in   legislature Self-government off agenda   BQ (11% of MPs)  from beginning of Martin premiership  (12 Dec 2003) to  Federal election  (28 Jun 2004)  Comparator party: New Democratic Party 5% of MPs & statewide party   Predicted behaviour: Using Parliament as a platform for grievances (Hypothesis 3)   BQ  (18% of MPs)  from Federal election (28 Jun 2004) to Fall of government (28 Nov 2005)   Comparator party: New Democratic Party 6% of MPs & statewide party   Predicted behaviour: Bargaining for parliamentary reform and regional benefits (Hypothesis 3) Figure 5.2 Paired cases, varying pivotal status; holding prospects for self-government constant  5.2 Participation in the central parliament. I use the online Hansard reports of proceedings in the Canadian and British Parliaments to gather data on the behaviour of the BQ, the SNP and their comparator parties. This chapter focuses on four types of parliamentary activity: contributions to no confidence debates, proposing Private Members’ Bills, tabling Private Members’ Motions and asking questions on upcoming parliamentary business.     In three of the four BQ cases, (1993-95, 1996-97 and 2003-04), the party is not pivotal. Given the government’s secure majority status and other parties’ lack of leverage, I look at parliamentary activities which are not subject to government agenda control. In the Canadian Parliament, for the periods under examination, there were no restrictions on MPs’ ability to table Private Members’  156 Bills or Motions.94  Most were not debated, but submitting these bills and motions to the official  record allows MPs to make a public statement about which issues they prioritize and which policy  or institutional reforms they favour. Table 5.1, below, sets out the number of Private Members’ Bills and Motions tabled in the House of Commons in each of my selected time periods. I examine two indicators of the BQ’s parliamentary behaviour. First, in order to test hypothesis 1 on disengagement,  I compare the number of bills and motions laid down by BQ Members to those submitted by  Members of comparator parties. Second, I focus on the subject matter of these bills and motions, looking for complaints about conditions in Québec (these are evidence the party is using a grieving strategy) or which seek institutional or constitutional reform other than changes to Québec’s  autonomy. These types of proposal would be evidence supporting hypothesis 3. They would show that the party is focused on the current central political system, not a new political system in a sovereign Québec.  Table 5.1 Total number of Private Members' Bills and Motions submitted in the Canadian Parliament Time period Self-government on the agenda? Status of BQ N Bills  N Motions 1993-95 Yes Non-pivotal 160 509 1996-7 No Non-pivotal 244 334 Feb-May 2004 No Non-pivotal 329 597 Oct 2004-05 No Pivotal 264 323 Source: Derived from data from the Parliament of Canada’s Status of House Business reports, 1996, 1997, 2004 and 2005.                                                    94 For the 35th Parliament, see Marleau and Montpetit (eds), House of Commons Procedure and Practice, 2000, Chapter 21 (Private Members’ Business) and for the 37th and 38th Parliaments, see the Annotated Standing Orders of the House of Commons, 2005, Chapter XI (Private Members’ Business).  Both are available on-line at: http://www.parl.gc.ca/ProceduralReference.aspx?Language=E There are restrictions on which motions and bills gain parliamentary time for debate and a vote. In the 37th and 38th Parliaments (discussed in Section 5.3) this was decided by lottery and in the 35th Parliament (discussed in Section 5.4) this was decided by a parliamentary committee (Loewen et al, 2014). The data collected for this chapter therefore focus only on the first readings of bills and introduction of motions. These are put on the Order Paper routinely, without taking up parliamentary time (see Commentary on Standing Order 86, Annotated Standing Orders of the House of Commons, 2005). Focusing on this initial stage means that the data are not distorted by the possible bias of committee proceedings against particular parties, individual MPs or topics (in the 35 th Parliament) or by the chance operation of the lottery (in the 37th and 38th Parliaments). Rather, they reflect only the actions of the MPs themselves in choosing whether to propose bills or motions on particular topics. 157 In the two earlier time periods (during the Chrétien government), the BQ is non-pivotal. So, the critical difference is the prospect of sovereignty for Québec. Before the referendum, the expected behaviour is disengagement, so we expect to see low numbers of BQ bills and motions, both in comparison to other parties and to the BQ itself in later time periods. After the referendum, we should see grieving: so we expect to see an increase in the BQ’s use of bills and motions and these should be focused on issues specific to Québec. In the two later time periods (the Martin government), sovereignty is off the agenda but the BQ’s legislative status changes. During the period of majority government (2003-04) the BQ should be grieving.95 During the period of minority government (2004-05) the Bloc should be bargaining for goods and services for the region and for institutional reform. In both these periods, the BQ has an incentive to use bills and motions so we do not expect to see a change in the frequency of use. However, their content should change. In both periods the BQ will focus on Québec’s concerns, but when it is pivotal there should be a greater focus on seeking parliamentary reform to favour small and regional parties.   Within my selected time periods, debates on explicit motions of no confidence in the government occurred twice, on March 23, 1977 and March 28, 1979.96 On both occasions, the SNP was pivotal. The very existence of the no-confidence motion shows the government’s vulnerability, indicating this is the prime time for other parties to extract concessions. Hypotheses 2 and 3 predict that pivotal  parties will attempt to bargain – but what they bargain for depends on whether self-government is imminent or not. We should therefore see differences between the SNP’s behaviour in the two                                                   95 Although there is a different Prime Minister, the BQ’s strategic position in 2003-04 is the same as in 1996-7. 96 These were in the UK. There were no such motions against the Canadian government during the case study periods of 1993-95 and 1996-7. There was a no confidence motion against the Martin government in the Parliament of 2004-2005, but this is not analysed here because there is no corresponding motion in its paired time period of 2003-04. 158 debates. In 1977, devolution was still a central part of the Callaghan government’s plans.97 So, according to hypothesis 2, the SNP should use the no confidence debate to reiterate its demands for devolution. By contrast, the 1979 debate was held after the no vote in the referendum on Scottish devolution.  In this debate we therefore expect the SNP to focus on its fortunes within the current political system. This priority can be expressed either as calls for institutional reform which would help small and regional parties such as themselves, or demands for benefits for their region, which will help their campaign in the next election (hypothesis 3).    Finally, I examine the SNP’s participation in the weekly questions on the “Business of the House,” where the Leader of the House sets out and justifies the agenda for the following week in the House  of Commons. Given the SNP’s pivotal position, this would be a forum where the party could state its demands and extract concessions. We are looking for two pieces of information. Firstly, does the  SNP participate in these sessions to a greater or lesser extent that the other small, pro-system parties, and does this increase once devolution is off the agenda? Secondly, if the SNP does participate, what demands does it make: does it use its leverage to demand legislative progress on self-government (hypothesis 2), parliamentary reform (hypothesis 3) or benefits for the region (hypothesis 3)?   5.3 The Bloc Québécois in the 35th Parliament: the impact of the sovereignty referendum For our purposes, the 35th Canadian Parliament falls into two periods: before and after the referendum on Québec sovereignty. Throughout this Parliament, the government had a secure majority. So, hypothesis 1 predicts disengagement by the Bloc when self-government was on the agenda, and grieving when it was not. This Parliament makes an interesting case study as it includes two new,                                                  97 Although the government had lost a guillotine motion at the committee stage of the Scotland and Wales Bill in February 1977, the Prime Minister had declared that a new devolution bill would be immediately introduced to Parliament. This was eventually passed in 1978. See Pilkington, 2002, pp 63-4 for an account of the passage of the 1970s devolution legislation for both Scotland and Wales.  159 essentially regional parties: the BQ itself and the Reform party whose seats were all in the Western provinces.98 Throughout this period, therefore, there were large numbers of MPs who were new to Parliament, and who had been elected on a platform which was critical of the existing political  system (Docherty, 1997). 99 The Reform party differed from the BQ, though, in that it aspired to  win seats throughout the country and form the federal government (Flanagan, 1995), a goal it later achieved.100 Furthermore, although critical of institutions such as the Senate and common practices  such as party discipline, the Reform party was not an anti-system party. Its goals were to reform the existing political system, not to destroy it or to leave it.   The data on Private Members' Bills, set out in Table 5.2, below, bear out the difference between the BQ and other parties. Compared to the BQ, Reform MPs made far greater use of Private Members' Bills, as did MPs in the New Democratic Party. The left-hand side of the table shows that before the sovereignty referendum, Reform MPs sponsored more than three times the number of bills sponsored by Bloc MPs. This gap was even starker for the NDP: its MPs sponsored 5 times the number of BQ Private Members’ Bills.101 This is despite the fact that the BQ was the largest opposition party. As Table 5.2 shows, before the referendum, the BQ supplied a smaller share of bills than would be expected, given its share of MPs, whereas both the Reform and NDP caucuses supplied more than would be expected.102 After the referendum, both Reform and NDP MPs sponsored three times the number of bills that the BQ did. The only party with comparable numbers to the BQ was the governing Liberal party, which included                                                   98 This statement is based on analysis of Parliament of Canada’s Members of the House of Commons, 1867-date database. The database is available at: http://www.parl.gc.ca/Parlinfo/Lists/Members.aspx?Language=E&Current=True  99 7 of the BQ’s 54 Members and 1 of the Reform Party’s 52 Members were incumbents (analysis of Members of the House of Commons, 1867-date database, ibid).  100 After merging with the Progressive Conservatives to form the Conservative Party of Canada.  101 These figures represent the average number of bills sponsored per MP, within each party in the parliamentary session. They are therefore standardized for the differences in the size of each party's caucus.  102 A chi-square test shows that before the referendum, there was a statistically significant (at the 99% level) gap between the observed cross-party distribution of Private Members’ Bills and the distribution that would be expected given each party’s seat share. The chi-square test statistic is 25.8780, with a p value of 0.000.   160 Table 5.2 Private Members' Bills in the 1993-97 Canadian Parliament   Jan 1994-Oct 1995 (Pre referendum)  Feb 1996-Apr 1997 (Post referendum)  Party % of MPs in Parliament103 % of all Bills Bills per MP % of MPs in Parliament104 % of all Bills Bills per MP Bloc Québécois (Official Opposition) 18.2% 11.9% 0.14 17.9% 10.7% 0.39 Liberal Party (Government) 60.1% 41.25% 0.14 60.5% 46.3% 0.50 Reform Party (3rd largest party; regional;  pro-system) 17.6% 35.6% 0.43 17.6% 36.5% 1.4 New Democratic Party (4th largest party; statewide; pro-system) 3% 11.25% 0.79 3% 6.1% 1.38  Note: The Bills per MP figures are standardized for the number of sitting days in the House of Commons in each time  period, using the following formula: ((N bills proposed / N MPs in party)/N sitting days in session)*100)    Source: Derived from data from the Parliament of Canada’s Status of House Business reports, 1996 and 1997.   Cabinet ministers who are not permitted to propose Private Members’ business (Marleau and Montpetit, 2000, Chapter 21). The data on Private Members' Motions also support the conclusion that BQ MPs participated less in parliamentary activities than other opposition parties (see Table 5.3, below).105 So, we know that the BQ was less engaged that other opposition parties throughout the 35th Parliament. According to my theory, we would expect this before the referendum, as the Bloc  would be focused on setting up a new political system in a sovereign Québec. However, after the                                                  103 Calculated on the basis of MPs sitting in the House at the beginning of the 1st Session of the 1993-97 Parliament.  104 Calculated on the basis of MPs sitting in the House at the beginning of the 2nd Session of the 1993-97 Parliament.  105 Similarly to the data on bills, a chi-square test shows that before the referendum there was a significant gap between the observed cross-party distribution of Private Members’ Motions and the expected distribution, given seat shares. The chi-square statistic is 257.2433 with a p value of 0.000. 161 Table 5.3 Private Members' Motions in the 1993-97 Canadian Parliament  Party Pre referendum  Post referendum   % of MPs in Parliament106 % of all Motions Motions per MP % of MPs in Parliament107 % of all Motions Motions per MP Bloc Québécois (Official Opposition) 18.2% 11.8% 0.43 17.9% 8.1% 0.41 Liberal Party (Government) 60.1% 24.6% 0.27 60.5% 19.9% 0.29 Reform Party (3rd largest party; regional; pro-system) 17.6% 23.4% 0.9 17.6% 38.6% 2.03 New Democratic Party (4th largest party; statewide; pro-system) 3% 40.3% 9 3% 33.4% 10.2 Note: The Motions per MP figures are standardized for the number of sitting days in the House of  Commons in each time period, using the following formula: ((N motions proposed / N MPs in party)/N  sitting days in session)*100)    Source: Derived from data from the Parliament of Canada’s Status of House Business reports, 1996 and 1997.   referendum failed, I predict that we should see the Bloc begin to engage in the current, Canadian political system, in order to justify why voters should send them to Ottawa. So, even if they still  engage less than other parties, we should see a change in the behaviour of Bloc MPs over time. The data on Private Members' Bills support this hypothesis. The Bills per MP figures in Table 5.2 (below) show that BQ MPs nearly trebled their use of Private Members' Bills after the referendum.108                                                   106 Calculated on the basis of MPs sitting in the House at the beginning of the 1st Session of the 1993-97 Parliament.  107 Calculated on the basis of MPs sitting in the House at the beginning of the 2nd Session of the 1993-97 Parliament.  162 The BQ is not alone in this trend. All parties increased their use of bills in the second session of the Parliament, but the Bloc still contributes a smaller proportion of bills than would be expected given  its share of MPs. However, in the post-referendum session, the gap between the parties’ share of bills  and the parties’ share of seats is no longer significant at the 95% level - in contrast to the pre-referendum session.109 Docherty’s research into the behaviour and attitudes of MPs during the 35th Parliament asked new Members how long it took them to become comfortable with the rules and procedures of the House of Commons. The mean response was 13.2 months (1997, p.89).110  Although this figure only covers Liberal and Reform MPs,111 it fits with the cross-party data here: all parties, including the BQ, show an increasing willingness to use the tool of Private Members’ Bills in the second session of the Parliament. This supports the hypothesis that after the referendum, the Bloc’s behaviour began to converge with that of other parties. Once sovereignty was off the immediate political agenda, its MPs began to engage with the existing Canadian institutions.  However, the data on Private Members' Motions show no such change: the figures for BQ MPs are almost exactly the same in the two time periods (see Table 5.3, above).112  Chi-square testing shows that the distribution of seats does not explain the cross-party distribution of motions, either before or after the referendum: Reform and NDP MPs contribute far more than would be expected.113  The BQ,                                                                                                                                                                        108 These figures are standardized to account for the number of days that Parliament was sitting in each time period. See the note for Tables 5.2 and 5.3. 109 The chi-square statistic is 9.8582 with a p value of 0.079. See footnote 103, above, for the comparable statistic for the pre-referendum period.  110 MPs were asked this question twice: as candidates during the 1993 election, they were asked to predict how long it would take them to learn rules and again in January 1996, at the beginning of the second session of the Parliament. As candidates, the mean response was 7.7 months. The figure quoted above is taken from the 1996 data, as these reflect MPs’ actual rather than anticipated experience (1997, 87-9).  111 The BQ’s response rate to Docherty’s surveys was too low to produce reliable results. See 1997, pp xvii-xviii.  112 The z statistic for the change in its share of motions between the two time periods is 76.8064 (p value 1.000). 113 Before the referendum, the chi-square statistic for the difference between the expected distribution of motions given party seat share, and the observed distribution of motions is 257.2433 (p value 0.000). After the referendum, it is 194.2936 (p value 0.000). 163 meanwhile, was under-represented in this particular parliamentary activity, with the referendum appearing to have no impact on its participation levels.114    Therefore, the bills data (but not the motions data) meet the prediction that BQ MPs will be less engaged in parliamentary activities when exit is an immediate prospect. However the theory states that once sovereignty is off the agenda, Bloc MPs will begin to participate in a specific manner. In this  Period, the party had no leverage with which to bargain, but did have to prepare for the next federal election. So, we should see the BQ begin to use the Canadian Parliament as a platform for Québec's concerns. Looking at the substance of Private Members’ Bills helps us to assess if this actually happened.  Table 5.4, below, shows the content of the BQ bills which were tabled before and after the referendum. These are categorized into bills which specifically concern Québec and/or language policy; state services provided to the whole country (e.g. Canada Post); the content or administration of the criminal law; regulation of the private sphere (e.g. amendments to the laws concerning personal finances, employment or taxation) and reform of central political institutions (e.g. to election laws or to the operation of the Canadian Parliament).  There is a clear shift in BQ MPs' priorities.115 Before the referendum, the Bloc mostly used bills to stake out a position on policy matters concerning the whole country: 94% of its bills concerned the criminal law, state services or the regulation of private relations. There were no BQ bills proposing changes to the Canadian Parliament or other central political                                                   114 One possible explanation for the Bloc’s increasing participation in bills, but not motions is the greater impact of a successful bill: it enters into law, whereas a motion simply endorses a statement or issue position. Loewen et al’s work on Private Members’ Bills and Motions in the Canadian Parliament shows that government backbench MPs who propose bills gain an electoral advantage (2014). However, this effect was not present for members of opposition parties such as the Bloc, nor did this research examine differences in the impact of bills and motions on electoral success. 115 The difference in the distribution of the BQ’s bills before and after the referendum has a chi-square statistic of 9.4697 (p value 0.050). 164 Table 5.4 BQ's Private Members' Bills by subject, 1993-97 Parliament Subject: Pre-referendum Post-referendum State services 29% 31% Private sphere 47% 10% Criminal law 18% 24% Quebec 6% 21% Institutional reform 0% 14% Source: Derived from data from the Parliament of Canada’s Status of House Business reports, 1996 & 1997  institutions. This reflects the prediction that the BQ would have short time horizons in the run-up to the referendum: why try to change a system it was actively trying to leave?   Only one of the BQ bills set out a specific demand concerning Québec, Québécois nationalism or the rights of the Francophone community (this was a pardon for Louis Riel). The BQ’s pre-referendum avoidance of Québec-specific issues is particularly stark when compared to other parties. Table 5.5, below, shows all the Private Members’ Bills devoted to Québec or Francophone issues in the 35th Parliament. Before the referendum, the proportion of such bills proposed by BQ MPs trails behind those proposed by Liberal MPs. The BQ’s share of these bills is even lower than that of the Western-based Reform Party. So, when sovereignty was on the agenda, the Bloc was not using the Canadian Parliament as a platform for expressing Québec’s grievances, which fits with the first hypothesis. In this period, BQ MPs were looking to a different forum altogether to make decisions about the region's services, economy and ethno-linguistic identity: a sovereign Québec Assembly.   After the referendum, there was a sharp increase in the BQ’s use of Private Members' Bills to address both Québec issues and institutional reform. Between 1996 and 1997, 14% of the BQ bills concerned institutional reform, suggesting the party’s MPs realized that they would now be participating in the central political system for the foreseeable future. Similarly, the proportion of BQ-sponsored bills   165 Table 5.5 Private Members' Bills on Québécois or Francophone issues, 1993-97 Parliament Proposed by: Pre-referendum Post-referendum BQ 8% 32% Liberals  (Government) 67% 36% Reform Party  (regional; pro-system) 17% 32% NDP (statewide; pro-system) 8% 0% Source: Derived from data in the Parliament of Canada’s Status of House Business reports, 1996 and 1997.   devoted to Québec's concerns rose from 6% pre-referendum to 21% post-referendum (see Table 5.4, above). This is also reflected in Table 5.5, showing all parties’ bills on Québec issues. From this we  see that after the referendum, the BQ proposed as many Québec or Francophone-related Private Members’ Bills as the Reform Party, and nearly as many as the Liberals.116 These bills included  changes to the Official Languages Act to increase the use of French in federal agencies, symbolic demands such as an official pardon for Louis Riel and attempts to increase Québec's autonomy on fiscal matters. These are the kind of activities the party can point to at the next election, when explaining voters should send sovereigntist MPs to the central parliament.   So, these shifts in the frequency and content of BQ Private Members’ Bills provide support for the  first hypothesis on disengagement. Once sovereignty was no longer an immediate prospect, the BQ began participating more in discussions about the Canadian parliament itself. The party also began to position itself as the Québec's champion in Ottawa, rather than solely a vehicle to achieve sovereignty.                                                     116 There is a significant (at the 95% level) difference between the cross-party distribution of Québec-related bills before the referendum and after it. The chi-square statistic for the difference in distributions is 8.3366 (p value 0.040).  166 5.4 The SNP and the Callaghan government: the impact of the devolution referendum The selected SNP time periods mirror the BQ cases in the previous section, as they fall either side of a referendum on increased autonomy for Scotland.117  However, the government’s minority status gave the SNP opportunities to extract concessions which were not open to the BQ. The second and third hypotheses on bargaining therefore come into play. In both periods, we should see the SNP engaging in parliamentary activities, because it is in a position to bargain. However, what the SNP bargains over should change. Before the referendum, we would expect the SNP to focus on the powers of the proposed Scottish Assembly, or on the referendum itself. After the referendum, we would expect it to bargain for institutional reform at Westminster and benefits for Scotland.   I focus on two month-long periods, when the SNP had maximum leverage. The first period, when devolution was on the agenda, begins on 24 February 1977. This was when the Conservative party gained a seat in a by-election, wiping out the Labour party’s parliamentary majority (Craig 1989). It ends with the Liberal-Labour pact, when the Liberal party agreed to support Labour on matters of confidence and supply, thus ensuring the survival of the government. The second period begins with the failure of the 1979 referendums on Scottish and Welsh devolution (on 1 March 1979) and  continues until the fall of the government on 28 March 1979. The Liberal-Labour pact had already ended (BBC, 1978), so the government was once again in the minority, making the SNP pivotal –  but devolution was now off the agenda. Both of these periods end with a no confidence motion, that the government won in 1977 but lost in 1979, precipitating the 1979 election.   Examining the weekly question periods on the business of the House of Commons shows that the                                                   117 Although in Scotland's case this was for the devolution of some legislative power within the existing territorial framework of the UK, not complete sovereignty. 167 SNP was engaged in routine parliamentary activities during both these periods (see Table 5.6, below). SNP Members asked questions of the Leader of the House as often as the pro-system Liberal party (which was of a similar size) and more often than the other regional parties in the House.118  Indeed, the SNP punched above its parliamentary weight: both before and after the referendum, its share of the questions was twice what would be expected given its seat share.119  Furthermore, there is no significant difference between the SNP's involvement in these sessions in 1977 and in 1979, suggesting that in both periods the party was involved in bargaining with the government of the day.120 So, as predicted, when the separatist party is pivotal, we see engagement in routine parliamentary activities regardless of whether self-government is on the agenda or not.  We also see this pattern in the no confidence debates. On both occasions, all SNP MPs voted against the government of the day, and the leader of the parliamentary party, Donald Stewart, made a speech setting out why the Scottish Nationalists would not support the government. This participation in extremely visible and high stakes debates shows that the SNP was not pursuing a disengagement strategy during either of these time periods. However, partaking in the more mundane day to day business of the House of Commons also shows that it was not simply using the legislature as a platform to express grievances and gain publicity. If the party was pursuing the latter strategy, we would see it                                                   118 Of these regional parties, the various Unionist parties were pro-system as they wanted Northern Ireland to remain part of the UK, whereas the Social Democratic and Labour Party favoured a united Ireland, so would be classified, like the SNP, as a separatist party. The position of Plaid Cymru is more ambiguous: throughout its history, and particularly in the 1970s it was ambivalent in its support for Welsh independence (see McAllister, 2001 for an overview). However, it always supported devolution of at least some powers to a Welsh legislature within the UK. It therefore opposed the constitutional status quo in 1977-79. 119 The z-statistic for the difference between the SNP’s expected question share (given its seat share), and the SNP’s actual question share is -12.1989 (p value 0.000) in the pre-referendum period. The corresponding z-statistic in the post-referendum period is -9.1815 (p value 0.0000). A z-test focusing on the SNP’s expected and observed share of questions was used rather than a chi-square test of the distribution of questions across all parties. The latter result would be largely a function of the comparison between the Labour and Conservative parties’ seats and questions share: as Table 5.6, shows, with seat shares of 49.8% and 44.3% respectively, they far outnumber the other parties. 120 The z-statistic for the difference between the SNP’s share of questions in 1977 and its share in 1979 is 3.6636 (p value 0.9999). 168 Table 5.6 All parties' use of questions on House of Commons business, 1977 and 1979, UK Parliament                       Source: Derived from data in Hansard, UK Parliament, 1977; 1979.  take part in prominent activities like questions to high profile Ministers or no confidence debates, but not questions on the forthcoming business of the House, which is unlikely to generate media coverage.   1977 (Pre referendum) 1979 (Post referendum) Party Number of Business Questions asked % of MPs in Parliament  % of all Business Questions asked Number of Business Questions asked % of MPs in Parliament % of all Business Questions asked Scottish National Party 5 1.7% 4.6% 4 1.7%   3.2% Labour Party (Government) 42 49.8% 38.5% 44 49.3% 34.9% Conservative Party (Official Opposition) 58 44.3% 53.2% 71 44.7% 56.3%  Liberal Party (statewide; pro-system) 4 2% 3.7% 5 2% 4% Ulster Unionist Party (regional; pro-system) 0 0.9% 0% 0 1.1% 0% Other Unionists (regional; pro-system)  0 4% 0% 0 0.5% 0%     Plaid Cymru 0 0.5% 0% 0 0.5% 0% SDLP 0 0.3% 0% 2 0.3% 1.6% TOTAL 109 100% 100% 126 100% 100% 169 So, if the SNP was engaging with the government of the day in order to bargain, what type of concession was it seeking, and did this change after the failure of the devolution referendum? The topics raised during House business questions, and the content of Stewart's speeches in the no confidence debates both point to the importance of Scottish self-government. Table 5.7, below, shows that, prior to the referendum, all but one of the SNP’s questions asked when the government would reintroduce the devolution legislation.121 This demand was explicitly linked to the SNP's threat to either introduce its own no confidence motion, or to vote against the government if any other party did so. For example on 24 February 1977, Margaret Bain MP asked:  “Will the right hon. Gentleman find time for an early opportunity to debate the Early-Day Motion in the names of my right hon. Friend the Member for Western Isles (Mr. Stewart) and the hon. Member for Carmarthen (Mr. Evans), since the question of continuing confidence in the Government is essential to the people of Scotland and Wales?”   The text of the Early-Day Motion in question is as follows: “That in view of their handling of the Scotland and Wales Bill, and their consequent loss of general credibility, this House has no confidence in Her Majesty's Government"  Hansard, 24 February 1977, Vol. 926, col. 1634.  In the 1977 no confidence debate we see the same focus on self-government. The SNP voted against the government (which survived with the help of Liberal votes). The Nationalists’ leader Donald Stewart used his speech to explain that in the SNP's view, Labour "most important" manifesto pledge had been to establish a Scottish Assembly (Hansard, 23 Mar 1977, Vol. 928, col. 1329). The government had therefore lost the SNP's support because the former had abandoned the devolution legislation in that parliamentary session.                                                     121 See footnote 97 for an overview of the passage of this bill.  170 Table 5.7 The SNP's questions on House of Commons business, 1977 and 1979, UK Parliament  1977 (Pre referendum) 1978 (Post referendum) Devolution / No confidence motions 4 2 Forestry 1 0 Fishing 0 1 Shipbuilding 0 1 Source: Derived from data in Hansard, UK Parliament, 1977; 1979.  Stewart also stated the price of future SNP support if the government survived: "I want that vote [on devolution] to be available to the Scottish people" Hansard, 23 Mar 1977, Vol 928, col. 1330  We therefore have evidence that prior to the referendum, the SNP was seeking to use its pivotal position to put pressure on the government to reintroduce legislation on devolution as quickly as possible. This is in line with hypothesis 2 which predicts how a separatist party behaves when it is pivotal and self-government is on the agenda. Hypothesis 3 states that once self-government is off  the agenda, a pivotal separatist party will use its leverage for regional benefits and parliamentary reform. However, there is only limited evidence of a change in strategy after the failure of the referendum in 1979. The number of House business questions which focused on devolution in 1979 was half that of 1977, and we do see a shift to demands for subsidies for Scottish industries (see  Table 5.7, above). This is consistent with being a regional champion in the central legislature: the party was seeking to use its pivotal status to extract additional government investment for Scotland. However, the numbers of questions involved are small and the difference is not significant at the 95% level.122   In addition, we do not see any questions focused on the SNP's procedural disadvantages in                                                   122 The chi-2 statistic for difference between the distribution of SNP question topics in 1977 and the distribution in 1979 is 0.9000 (p value 0.343).  171 Parliament, nor requests for institutional reform. This is contrary to the third hypothesis that when self-government is off the agenda, a separatist party will use its pivotal status to increase its powers and resources within the central legislature. Furthermore, the speech made by Donald Stewart in the 1979 no confidence debate was strikingly similar to that made in 1977. It again focused on devolution, blaming the government for the referendum's failure.123 Stewart anticipates the election of Thatcher's Conservative government at the coming election and turns his attention to her:  “The problem [self-government for Scotland] will not go away. It will be lodged on her desk  and will have to be dealt with.” Hansard, 28 Mar 1979, Vol. 965, col. 490  This last quote suggests that in the eyes of the SNP leadership, devolution was not off the agenda; despite the failure of the referendum, it believed that it would be able to resurrect the issue under a future government. If that was the case, then the party’s continued use of the Westminster Parliament  to make demands of both the government and the official opposition suggest that it was still using a bargaining strategy, as per hypothesis 2. However, we don’t find support for hypothesis 3 – that once self-government is off the agenda, separatists will use pivotal status to bargain for regional benefits  and institutional reform. Instead, the SNP’s continued emphasis on self-government, its silence on parliamentary reform and the low number of demands for subsidies for Scottish industry suggest that the SNP leadership was still focused on a hypothetical Scottish political system, not the existing British one.                                                   123 Stewart argued that the vote threshold for adopting devolution was illegitimate (Hansard, 28 Mar 1979, Vol 965, col. 489). This required 40% of the registered electorate (not those actually voting) to vote in favour, thus tying the outcome to turnout.  172 5.5 The Bloc Québécois in the 37th and 38th Parliaments: the impact of minority government. The 2004 Canadian election reduced the Liberal government to minority status in the House of Commons, while the BQ increased its share of MPs from 11% to 18%. This gave the BQ a new  source of leverage with which to demand concessions from the government – the first time it had  been in this position since it was formed in 1990. Hypothesis 3 predicts that we should see the party switch from a grieving to a bargaining strategy. It should be engaged in parliamentary activities  in both periods (because self-government is not on the agenda). However, when pivotal the party should be more interested in parliamentary reform, as it actually has the opportunity to gain concessions.   The data on Private Members’ Bills and Motions do not support this thesis. In fact they show that  the BQ was disengaged from these activities both when the Martin government had a majority  and when it did not. Tables 5.8 and 5.9 (below) show that BQ MPs made far less use of bills and motions than other opposition Members, and significantly less than would be predicted by the  BQ’s proportion of MPs.124  The BQ’s use of bills and motions actually fell substantially during the period of minority government, when the theory’s prediction was that it would remain stable. However, this latter finding should be treated with some caution. Tables 5.8 and 5.9 show that there was a similar drop in the use of bills and motions by all parties in the 38th Parliament, suggesting that there is a  systematic institutional factor at work rather than a simple change in the behaviour of BQ MPs  alone.  When the BQ’s proportion of all parties’ Private Members’ Bills and Motions are compared between the majority government and minority government periods, the difference is not significant                                                   124 Using a chi-square test to compare the observed cross-party distribution of Private Members’ Bills with the distribution that would be expected based on each party’s seat share, shows differences significant at the 99% level in both periods. For the majority government period, the chi-square test statistic was 95.3054, with a p value of 0.000. For the minority government period, the chi-square test statistic was 119.7034 with a p value of 0.000. Performing the same analysis for Private Members’ Motions yields a chi-square test statistic of 282.7098 (p value 0.000) during the majority government period and a chi-square test statistic of 205.5932 (p value 0.000) during the minority government period. 173 Table 5.8 Private Members' Bills in the Canadian Parliament, 2004-05.   Feb-May 2004 (Majority government)  Oct 2004- Nov 2005 (Minority government)  Party % of MPs in Parliament125 % of all Bills Bills per MP % of MPs in Parliament126 % of Bills Bills per MP Bloc Québécois (3rd largest party) 11.3% 7.6% 1.34 17.5% 8% 0.24 Liberal Party (Government) 57% 25.2% 0.88 43.2% 10.2% 0.13 Conservative Party of Canada (Official Opposition) 24.3% 42.6% 3.49  31.8% 49.6% 0.84 New Democratic Party (4th largest party; statewide; pro-system) 4.7% 21% 8.96 5.8% 31.1% 2.85 Note: The Bills per MP figures are standardized for the number of sitting days in the House of Commons in each time period, using the following formula: ((N bills proposed in session / N MPs in party)/N sitting days in session)*100)    Source: Derived from data from the Parliament of Canada’s Status of House Business reports, 2004 and 2005.   at the 95% level.127  Nevertheless, the cross-party comparison does not support the idea that the BQ Members were pro-actively raising grievances or proposing institutional reform in either parliamentary session.                                                     125 Calculated on the basis of MPs sitting in the House at the beginning of the 3rd Session of the 2000-04 Parliament.  126 Calculated on the basis of MPs sitting in the House at the beginning of the only session of the 2004-05 Parliament. 127 When comparing the BQ’s proportion of bills in the majority government (T1) and the minority government (T2), the z-statistic for the difference is 32.9205 (p value 1.000). In calculating this statistic, the mean for T2 was multiplied by BQ’s share of seats at T1/BQ’s share of seats at T2, in order to standardize the BQ’s caucus size across the two time periods. For motions, the z-statistic for the difference between T1 and T2 is 37.2474 (p value 1.000), controlling for the BQ’s caucus size in the same way as the bills calculation. 174 Table 5.9 Private Members' Motions in the Canadian Parliament, 2004-05   Feb-May 2004 (Majority government)  Oct 2004- Nov 2005 (Minority government)  Party % of MPs in Parliament128 % of all Motions Motions per MP % of MPs in Parliament129 % of Motions Motions per MP Bloc Québécois (3rd largest party) 11.3% 4% 1.21 17.5% 6.5% 0.24 Liberal Party (Government) 57% 20.8% 1.26 43.2% 11.8% 0.18 Conservative Party of Canada (Official Opposition) 24.3% 37.5% 5.5 31.8% 24.1% 0.5 New Democratic Party (4th largest party statewide; pro-system) 4.7% 35.3% 27.4 5.8% 53.6% 6.01 Note: The Motions per MP figures are standardized for the number of sitting days in the House of Commons in each  time period, using the following formula: ((N motions proposed / N MPs in party)/N sitting days in session )*100)    Source: Derived from data from the Parliament of Canada’s Status of House Business reports, 2004 and 2005.    The second element of hypothesis 3 on bargaining concerns the content of the BQ’s motions and  bills – as this shows us which issues BQ MPs are prioritizing. Tables 5.10 and 5.11, below, break  down the content of the bills and motions that were tabled by BQ Members. These tables use the  same coding scheme as section 5.3: Québec and/or language policy; state services provided to the whole country; the content or administration of the criminal law; regulation of the private sphere  and reform of central political institutions (e.g. to election laws or to the operation of Parliament).  Here, there is some evidence to support the thesis that separatist parties will use pivotal status to                                                   128 Calculated on the basis of MPs sitting in the House at the beginning of the 3rd Session of the 2000-04 Parliament.  129 Calculated on the basis of MPs sitting in the House at the beginning of the only session of the 2004-05 Parliament. 175 Table 5.10 BQ Private Members' Bills by subject, 2000-04 and 2004-06 Parliaments Subject: Majority government Minority government State services 28% 24% Private sphere 60% 48% Criminal law 12% 14% Québec 0% 0% Institutional reform 0% 14% Source: Derived from data from the Parliament of Canada’s Status of House Business reports, 2004 and 2005.  Table 5.11 BQ Private Members' Motions by subject, 2000-04 and 2004-06 Parliaments Subject: Majority government Minority government State services 46% 33% Private sphere 29% 62% Criminal law 0% 0% Québec 21% 0% Institutional reform 4% 5% Source: Derived from data from the Parliament of Canada’s Status of House Business reports, 2004 and 2005.  seek institutional reforms that favour small and regional parties. When the BQ had no leverage, its MPs did not table any bills seeking institutional reform, but during the period of minority government, 14% of their bills concerned this subject. These bills concerned the administration of federal elections (indicating that BQ MPs anticipated running in future elections) and the role of Parliament itself (specifically concerning the powers of Officers of Parliament and Parliament’s ratification of international treaties).  However, two aspects of these data suggest that the BQ took only limited advantage of minority government to seek these kind of concessions. Firstly, the same pattern is not seen in BQ motions.  As Table 5.11, above, shows, the proportion of the party’s motions devoted to institutional reform remained static over the two time periods.130 Secondly, Tables 5.12 and 5.13, below, demonstrate  that, compared to the other opposition parties, the BQ played a very limited role in suggesting                                                   130 Similarly, when the BQ’s share of all parties’ motions on institutional reform is compared across the two time periods, there is no significant difference. The z-statistic for the difference is 1.4052 (p value 0.9200).  176 Table 5.12 Private Members' Bills on institutional reform, 2000-04 and 2005-06 Parliaments Proposed by: Majority government Minority government BQ 0% 9% Liberals (Government) 29% 9% Conservative Party of Canada (Official Opposition) 44% 52% NDP (4th largest party; statewide; pro-system) 21% 21% Source: Derived from data from the Parliament of Canada’s Status of House Business reports, 2004 and 2005.  Table 5.13 Private Members' Motions on institutional reform, 2000-04 and 2005-06 Parliaments Proposed by: Majority government Minority government BQ 2% 4% Liberals (Government) 7% 17% Conservative Party of Canada (Official Opposition) 64% 42% NDP (4th largest party; nationwide; pro-system) 23% 29% Source: Derived from data from the Parliament of Canada’s Status of House Business reports, 2004 and 2005.  reform to Parliament or the electoral system. The Official Opposition, the Conservative Party of Canada, proposed far more reforms than any other party. Similarly, during Paul Martin’s minority government, the NDP proposed more than twice the number of BQ reform bills and six times the number of BQ reform motions, despite having a caucus which was a third the size of the BQ. Chi-square testing confirms that there was no significant change in the party distribution of institutional reform bills between the majority and minority time periods.131                                                   131 The chi-square statistic for the party distribution of institutional reform bills during majority government and during minority government is 3.7210 (p value 0.445).  177 Furthermore, there is no compelling evidence that the BQ used its leverage to act as a regional champion for Québec, seeking substantive benefits for the province. No BQ Member tabled a bill  related to Québec in either time period (see Table 5.10, above). There were 5 motions sponsored by BQ Members in the 37th Parliament which concerned either Québec or the wider Francophone  community. However, contrary to hypothesis 3 on bargaining, there were no such motions in the  38th Parliament, when the party was pivotal (see Table 5.11, above). In this, the BQ was no different to the other parties. Analysis of all parties’ Private Members’ Bills and Motions in these two periods  suggests that specific concessions to Québec, or discussion of the region’s grievances rarely figured  on the parliamentary agenda at all.132 During the period of minority government, there were no  motions and only one bill on Québécois or Francophone issues from any party.133 During the period  of majority government, there were three bills and seven motions proposed on Québécois or  Francophone issues. Of these, five motions (as already stated) were put down by BQ MPs, one by a Progressive Conservative and one by an Independent. The three bills were tabled by Liberal and Independent MPs.   So, we do not see the BQ using its pivotal status to put Québec issues on the agenda. Furthermore,  Bloc MPs only played a minimal role in proposing reform to the central political system. However, during a period of minority government, opposition parties have a greater incentive to co-operate  with one another, as there is a real chance they may successfully pass a motion or bill which has  cross-party support. To test this possibility, I recorded all motions and bills that were proposed and seconded by MPs from different parties. The results do not suggest that the Bloc played a supporting role in other parties’ attempts to gain concessions for Québec or achieve institutional reform. Bloc MPs                                                  132 As discussed in section 5.3, this is in contrast to the earlier Parliament of 1993-97, when the referendum on sovereignty made Québécois grievances salient to all parties.  133 The bill was tabled by a NDP MP from Manitoba, and as it concerned a pardon for Louis Riel was on a topic that concerned that province as much as it did Québec.  178 did not second any bills or motions on either of these subjects, in either time period. Furthermore, of the party’s own bills and motions on these topics, only one attracted a cross-party seconder,134 and this was during the period of majority government, when we would not predict cross-party bargaining. All other instances of BQ co-operation with other parties concerned either Canada-wide policy135 or international issues136 and involved a NDP Member seconding a Bloc motion or bill. This suggests that any inter-party co-operation was based on the two parties’ similar left-wing ideology137 rather than on the Bloc’s goal of sovereignty. Furthermore, this BQ-NDP co-operation occurred more frequently in the period of majority government than in the period of minority government.   In sum, therefore, there is little to suggest that pivotal status affected either the BQ MPs’ participation in bills and motions, or its interaction with other parties. This means that there is no evidence for hypothesis 3, on bargaining when self-government is off the agenda.  5.6 Conclusion Taken together, the BQ and SNP cases offer support for two of the hypotheses on separatist behaviour. However, contrary to hypothesis three, pivotal status was not as important as predicted – at least, not when self-government was off the agenda. The transition from majority to minority government in                                                  134 This was M-238, which condemned the 18th century deportations of the Acadian (French-speaking) community. This was proposed by a Bloc MP during the 3rd session of the 37th Parliament and seconded by a Conservative MP (Parliament of Canada, 2004).  135 These comprise 6 bills in the 37th Parliament, on the Labour Code, national anti-poverty initiatives, old age security and the decriminalisation of prostitution and 3 bills in the 38th Parliament on income tax, excise duties and national economic development (Parliament of Canada 2004; 2005). 136 These comprise 1 bill in the 37th Parliament on a Free Trade Agreement with Israel, 1 motion in the 37th Parliament on the Armenian genocide and 1 motion in the 38th Parliament on the World Trade Organization (Parliament of Canada 2004; 2005).  137 When placing Canadian parties from this era on a left-right spectrum, scholars have labelled the NDP as the most left-wing party, followed by the BQ, with the Liberals in the centre and the Reform/Alliance/Conservative Party of Canada as the most right-wing party. See, for example, Cross and Young’s findings on the ideological position of each party’s membership (2004) and Nevitte et al’s findings on the positioning of each party’s supporters in both the 1993 and 1997 elections (1999, 52-54). 179 2004 appears to have made little difference to Bloc MPs’ use of motions and bills, despite this being the most direct way for opposition MPs to put their preferred issues on the agenda.   Conversely, there is evidence that the prospect of self-government does have an impact. In line with hypothesis 1, the BQ displayed clear signs of disengagement when sovereignty was on the agenda and it was not pivotal. This is true both when comparing Bloc MPs’ behaviour to other opposition parties and to their own actions after the referendum had failed. Data on Bloc MPs’ use of Private Members’ Bills fit with the prediction of a grievance strategy once self-government is no longer imminent. However, the data on Private Members’ Motions do not show the same pattern. This is puzzling: although motions do not become law, in reality, Private Members’ Bills are very unlikely to do so either, particularly when proposed by opposition parties that are not pivotal. Both measures, therefore, can be thought of as a mechanism for the MP to state her policy interests and gain publicity rather than an attempt to create substantive change.   The pivotal SNP, meanwhile, showed sustained participation in both high profile and routine parliamentary activities, before and after the referendum. Before the referendum, the SNP used its position to bargain for concessions on the devolution referendum, as predicted. When that was no longer an option, there were some limited signs that the party had shifted to a focus on regional investment. However, in the most crucial (and visible) debates it continued to prioritize its primary goal of self-government, contrary to hypothesis 3.  So, similarly to the BQ, pivotal status did not have the expected impact when self-government is off the agenda. However, the interaction of self-government being on the agenda and pivotal status does lead to bargaining on self-government – as predicted by hypothesis 2.    180 The data examined here allow us to draw conclusions of wider use to the conflict management and institutional literature. Time horizons did appear to alter separatist behaviour: there was a clear shift in the BQ’s behaviour once the sovereignty referendum had failed. However, as the SNP case shows, parties have to accept that self-government is truly off the agenda in order to switch to secondary concerns such as winning seats in the next election and parliamentary reform. From a normative perspective, these findings also undermine the concerns of scholars of anti-system parties. Separatist parties were not always obstructive or disengaged: like all parties, their behaviour changed in response to their immediate priorities and their opportunities to pursue those goals. Finally, both separatist parties were able to use parliamentary activities to advocate and bargain for self-government. This suggests that Canadian and British institutions have a certain resilience: providing space for attacks on the political system without suffering long-term harm.      181 Chapter 6: Conclusion In this chapter I review my findings about the behaviour of the Scottish National Party, the Bloc Québécois and the Irish Parliamentary Party, asking whether there is evidence to support the hypotheses laid out at the end of Chapter 2. Taken together, these conclusions increase our understanding of three political phenomena: anti-system parties, the use of political institutions for conflict management and small parties’ participation in parliamentary activities. I then consider the limitations of the research presented in this dissertation in order to set out a future research agenda on separatist parties.  6.1 Key findings Of the three independent variables, the most important for predicting separatist behaviour was the prospect of self-government. Chapter 3 shows that the Scottish, Irish and Québécois nationalists all focused their campaigns on anti-system grieving – when some form of devolution, sovereignty or outright independence was on the political agenda. The hardest test for this theory was the IPP in the 1886 election, as all the other independent variables created incentives for the party to use a credit-claiming strategy. Instead, the party ignored the fact that it had a pivotal position in the British Parliament and co-operative relations with the incumbent government. It focused on persuading its supporters that Ireland could never prosper in a British system, in order to shore up support for Irish Home Rule.   However, once the chances of self-government recede, we see a statistically significant shift in electoral rhetoric. All the parties examined switch to explaining their role within the current political system. They emphasize the public expenditure and policy reforms that their parliamentary work has 182 secured for their region (the IPP in 1895), and they attack the government of the day, rather than the constitutional set-up (the SNP in 1979 and 2005 and the BQ in 1997).   The prospect of self-government also has an impact on legislative behaviour. Chapter 5 shows that the Bloc MPs change their approach part-way through the 35th Canadian Parliament. After the failure of the 1995 referendum on sovereignty, they sharply increase their use of Private Members’ Bills. In particular, they start to use these Bills to put issues of concern to Québec and the wider Francophone community on the parliamentary agenda. This foreshadows the change we see between the party’s 1993 and 1997 electoral campaigns. In the former, a vote for the Bloc is portrayed as a vote in support of sovereignty. In the later campaign, the Bloc was instead explaining what it could achieve for Québec within the Canadian Parliament, and indeed, its MPs had been making greater use of parliamentary tools to bring attention to Québec’s grievances.   The second independent variable is electoral competition: whether separatists are competing for seats with the government of the day or an opposition party. Chapters 3 and 4 show that this has a powerful conditional effect. It makes no difference to electoral rhetoric when self-government is on the agenda: this is because the separatist party is focused on its primary goal of persuading citizens to support constitutional change. However, when such reforms are a distant prospect, or when they have recently failed, the separatists I examine are pragmatic. They focus on winning seats within the central legislature, in order to remain relevant. The identity of their electoral competitor therefore becomes crucial, and shapes their appeals to voters. When competing against the government of the day, separatists employ a grievance-based strategy which points out the region’s problems and blames the incumbent ministers and their policy choices. In other words, separatists act in the same way that 183 mainstream, pro-system opposition parties do. We therefore see the IPP attack the Conservative government in 1906.   The other cases show that this is not simply a matter of ideological disagreement. The SNP has left-of centre policies yet reserves its harshest criticism for the Labour governments of 1979, 2001 and 2005. In 1983, it even attacks the Labour government that had been ousted four years previously, rather than the clearly right-wing Thatcher government in power at the time. The Bloc, meanwhile, attacks whichever party is the biggest threat, regardless of ideological proximity: in 2004 and 2006 this is the incumbent Liberal government, while in 2008 it is the incumbent Conservative government.   However, when separatists are competing with the opposition party, we see a very different electoral strategy. The IPP in the 1909 and 1910 elections and the SNP in 1983 include credit-claiming in their appeals. The parties argue that their presence in the central parliament puts pressure on the government, resulting in benefits for their region (for example land reform and Catholic schools for the Irish or increased military pensions for the Scottish). Indeed, in the 1979 election we see that the SNP switches electoral rhetoric between individual constituencies: using credit-claiming where the opposition party is their main rival and using grievances where they compete against the government. This shows that only two months after the failed 1979 referendum on Scottish devolution, the SNP had already turned its mind to detailed electoral calculations within the British system.   The weakest of the three independent variables is pivotal status. Chapter 4 shows that neither the IPP in December 1910 nor the Bloc in 2006 and 2008 shift their electoral rhetoric to reflect the minority government of the previous Parliaments. This fits with the predictions made in Chapter 2: it is the structure of electoral competition which offers the greater opportunities for separatist parties to gain 184 concessions, and to claim credit for them. This is partly because of the incentives to the government to make such concessions - separatists taking seats from the opposition is good news for the government. It is also due to the separatists’ own electoral calculations: if the separatists are competing with the government, they do not want to acknowledge that the latter has done anything useful for the region. So, credit-claiming will be rejected even if the separatist party was pivotal and could have extracted concessions.   Finally, pivotal status has limited impact even in the legislature itself. Chapter 5 shows that under minority governments, Bloc MPs did not significantly change their use of Private Members’ Bills and Motions. So, as mentioned above, the prospect of self-government makes a difference to separatists’ legislative participation, but actual parliamentary leverage does not. Indeed, the only impact that pivotal status does seem to have is indirect. When self-government is on the agenda and separatists are not pivotal, they disengage from the central legislature (see the BQ in 1993-95). This makes sense, as they don’t expect to be participating in that institution over the long term. However, the SNP in the 1970s shows that a separatist party will participate in central parliament when self-government is on the agenda, when it is pivotal. This again makes sense, as the separatists can use their leverage for their primary goal of self-government, for example by supporting relevant legislation and by bargaining over the conduct of constitutional referendums. So, pivotal status does not alter the separatist party’s legislative objectives, nor its campaign strategy – but it does alter the opportunities open to the party. How the party uses that leverage depends on whether it is focused on an independent political system, or the institutions it is currently in.   185 6.2 Contributions These findings add to our understanding of anti-system parties, conflict management and the activities of small parties. First, the IPP, SNP and BQ’s behaviour undermines the concerns of scholars of anti-system parties. Sartori and Schedler argue that these parties will either disengage entirely from democratic institutions or seek to disrupt legislative activities (1971; 1996). They state that this undermines the legitimacy of the political system, as voters either see chaotic decision-making in the parliament or a significant section of voters going unrepresented. Furthermore, Keren argues that this de-legitimization is reinforced by the rhetoric of the anti-system parties themselves. As these parties are not involved in the compromise of government (nor do they expect to be) they can preach a form of “political perfectionism” which sets up unrealistic expectations of a new political order – that the existing system can never live up to (2000). This, matched with legislatures which have been boycotted or disrupted, undermines the electorate’s faith in existing electoral and legislative institutions, and ultimately the democratic process.  So, according to these scholars, the behaviour of anti-system parties will be rigid and flows directly from their objectives: they reject the legitimacy of the current system, so will constantly attack it. Young and Bélanger add more nuance to this picture, by pointing to multiple examples of separatist parties that do engage in the legislature (2008).  However, their categorization is at the party level, distinguishing between “democratic” and “obstructionist” separatists. This still assumes that the behaviour of each party is static over time, and therefore does not respond to external stimuli.  My research complicates this picture, showing that separatist parties do not constantly undermine the political system, and, crucially, the same party shifts its rhetoric over time.  The case of the Bloc in the 35th Canadian Parliament shows that the exact same set of MPs can significantly shift their behaviour within a single electoral cycle. They did this when their time horizons changed. After the loss of the 186 1995 referendum, these MPs expected to be running in another federal election and if successful, sitting in the Ottawa Parliament for at least another four years. At this point, we see Bloc MPs making much greater use of parliamentary tools available to them, in order to publicize the grievances of their region and linguistic community.    We also see separatists varying their electoral rhetoric. As Sartori et al would predict, there is much use of anti-system grieving, coupled with utopian claims about the potential of a self-governing polity. However, when self-government is off the political agenda, we see that separatists actually do acknowledge that their region has received concrete benefits under the current constitutional arrangements. Furthermore, they do this in statewide campaigns – during high profile events when they know their message will be relayed to voters. The fact that we find this credit-claiming in manifestos shows that this is a deliberate rhetorical choice, sanctioned by the party leadership. This runs contrary to the predictions of the anti-system party literature, as it shows that separatists will, under certain circumstances, use rhetoric which is actually pro-system.   Finally, direct participation does not only happen when separatists expect to be in the current system for the foreseeable future. In 1977, when Scottish devolution was on the agenda, the SNP still engaged in routine parliamentary business, discussing the upcoming agenda for the British House of Commons. This was not an activity that got newspaper coverage, so was not about seeking publicity. Rather, it showed that separatists will actually use a central institution to pursue their own agenda of exit from that very institution.   However, this engagement is not random – Chapters 3-5 show that we can predict when separatists will engage and when they will disengage, when they will attack the overall system, when they will attack 187 the government of the day and when they will use their position to extract concessions for their region. It is a matter of opportunities: separatists are more likely to engage in routine parliamentary activities when they are pivotal and can therefore bargain. They are more likely to claim credit for spending in their region when the government of the day has incentives to help them do so. Finally, when self-government is off the agenda, separatists become very concerned with electoral competition, focusing their campaigns on attacks against the party most likely to take seats from them. In other words, as separatists’ time horizons increase, their legislative and electoral behaviour looks increasingly similar to that of mainstream parties.  The conflict management literature raises opposite questions to those of the scholars of anti-system parties. Armed separatist movements are the largest cause of ethnic conflicts in the post-war era (Harbom et al, 2008). In an earlier period, the constitutionalist IPP was constantly challenged by more radical Irish Nationalist groups who advocated violence over political negotiation. Self-government was eventually decided through a war of independence and civil war within Ireland itself, as opposed to electoral campaigns and parliamentary bargaining. This raises the question of whether it is possible for separatist parties to achieve anything within the democratic process: if they cannot, secessionist sentiment may turn to illegal or violent means instead. So, whereas the literature on anti-system parties asks if separatist parties damage the democratic process, conflict management scholars ask whether democratic institution can, or should, respond to separatist sentiment, in order to keep the peace.  On one side of the debate is empirical research suggesting that accommodating one separatist movement encourages the emergence of other secessionist groups and encourages these groups to actively rebel against the state (Brancati, 2006; Walter, 2006b). Sorens has challenged this work, 188 arguing that “quasi-legal”138 routes to separation make armed regional rebellion less likely. Furthermore, he argues that if the central government can credibly commit to make concessions to a region, it lowers the costs and increases the benefits of the status quo, relative to self-government – so reducing popular support for separation (2012).   For these scholars, “successful” conflict management is not necessarily about preventing secession, but about avoiding violence and minimizing legal, political and economic uncertainty. In particular, there is a concern that blocking certain viewpoints or objectives from democratic institutions makes it more likely that these groups will turn instead to extra-constitutional, sometimes violent outlets. (Crick, 1971; Sinpeng, 2013; Sorens, 2012). From this viewpoint, it is encouraging that the democratic institutions in this research did provide a forum for discussing controversial and divisive issues – even the break-up of the country. The IPP and the SNP both used routine parliamentary activities and high profile debates to raise the issue of self-government, and the BQ has introduced motions to debate outright sovereignty (see Chapters 3 and 5). Furthermore, when self-government was off the agenda, separatist parties were able to win benefits for their regions, and tell voters about these successes (see Chapter 4).  So, in advanced democracies at least, even those who rejected the democratic legitimacy of an institution were able to participate in it and, crucially, gain both constitutional and distributive concessions.   A final contribution of this research is to research on small parties in legislative institutions. Chapter 4 demonstrates the importance of the “electoral competitor” variable. It also shows that the separatists focused their grievances on the party most likely to take its seats (or from which it was most likely to gain seats). This focus on seat count, as opposed to vote share, suggests that separatists do derive utility                                                  138 Defined as the state stating clearly that it will not use military or legal coercion to prevent separation. 189 from participating in central institutions. The inference fits with recent cross-national research by Dinas et al which shows that small parties obtaining parliamentary seats had a subsequent electoral advantage over those that did not (2014). The authors theorize that this may be due to influence over policy, increased media visibility or because seats signal electoral credibility to sympathetic voters, preventing strategic desertion to larger parties.   The research here casts light on this debate, at least for separatist parties. We see that pivotal status has very little impact on legislative activities, or electoral rhetoric, when self-government is off the agenda. The Bloc, for example, does not take advantage of its leverage in the 2004-2011 period by getting more involved in Private Members’ Motions or Bills, nor by changing the issues that it focuses on. The IPP uses credit-claiming in its electoral campaigns even when it had no leverage in the prior parliament. Conversely, the SNP did not use credit-claiming when it was pivotal.   However, separatists do use pivotal status when self-government is on the agenda in order to pressure governments to deliver their promises. In the 1886 campaign, we see IPP candidates openly talking to their voters about how they used pivotal status to extract the Home Rule Bill from Gladstone – and planning to continue that campaign if pivotal in the next parliament (whichever party was in power). Nearly a century later, the SNP uses both routine questions on parliamentary business and set-piece no-confidence debates to demand legislation on Scottish devolution. Again, this demand is directed to both the incumbent Labour government and the (assumed) incoming Conservative government. This suggests that for separatists, parliamentary status is valued for policy influence in one issue area only: self-government. At the same time, we have evidence that electoral competition also drives separatist behaviour - when self-government is off the agenda. This suggests that, following Dinas et al (2014), parliamentary status is valuable for policy reasons when a small party’s signature issue is already on 190 the agenda (and therefore, change is plausible). However, the chance to participate in parliamentary activities does not appear to be valued in other, secondary policy issue areas. When self-government is off the agenda, parliamentary status is valuable to separatists for instrumental reasons: increased media visibility and a signal that support for the party is not a wasted vote. In summary, separatist parties pursue self-government when it is plausible, but when it is not, they work hard to maintain their electoral position in the existing system.  6.3 Limitations of the research The theory and predictions laid out in Chapter 2 are based on a simplified model of a parliamentary system with a Single Member Plurality electoral system, two country-wide, mainstream parties and one separatist party. Accordingly, there are limits to the generalizability of the findings in this dissertation.  First, the independent variable “electoral competitor” assumes two mainstream parties (or at least blocks of parties) which alternate in government. Therefore, if governments lose seats to separatists, it benefits the official opposition party and vice versa. In multi-party systems, it is not a straightforward zero-sum game – so it may not be as clear to governments when they can safely co-operate with separatists. Nor will it be obvious to the separatist leadership when they should grieve against the government of the day, or when they can credit-claim without helping an electoral rival. We already see this problem in Chapter 4, where several of the Canadian federal elections are difficult to categorize on this independent variable. In Canada, regional cleavages (not only in Québec) cut across the left-right spectrum, meaning that since 1993 there have been five parties who were plausible contenders for government or official opposition status, or which have been pivotal. This meant, for example, that in the run-up to the 2011 election, it was not clear to observers (or presumably the Bloc leadership) which 191 party was the Bloc’s primary competitor for seats. This makes it difficult to predict the campaign strategy in elections when self-government is off the agenda.  Similarly, more proportional electoral systems will make it difficult to predict separatists’ electoral behaviour. These systems will increase the number of potential competitors for separatist seats (Cox, 1997; Duverger, 1954). Proportional representation which distributes seats statewide will take away the separatists’ advantage of regional concentration (Meguid, 2008). However, a regional tier of seats might advantage the party, depending on whether the boundaries of the regional tier coincide with the separatists’ ethno-national group. In either case, the party would be potentially competing against all other parties, rather than one easily identifiable rival. Preferential voting, on the other hand, would sharpen the dilemma about appealing to separatist voters, protest voters or benefit-seeking voters. Separatists would be seeking second or third preferences from voters who supported other parties. However, if they had built a credit-claiming strategy in order to attract the first-preferences of benefit-seeking voters, this would undermine the chances of attracting the later-preferences of protest voters and vice versa. The theory laid out here does not give us insight into how the leadership would resolve these problems.  A second limitation is that the theory concentrates on the motivations and actions of the separatist party. The government’s actions are theorized when discussing credit-claiming, to predict whether there will actually be any regional investment that the separatists can point to in their campaign. However, the theory does not consider whether and how the government will make concessions on self-government or on institutional reforms within the current system. The actions of the opposition party are not discussed either, other than to say that they will try to target protest voters who are angry with the incumbent government. In considering the interaction of parties, Meguid’s work is useful as 192 she focuses on how mainstream parties respond to the new issues raised by niche parties: setting out when large parties will try to co-opt a niche party’s position, when they will ignore the new issue and when they will take up a policy position at the opposite end of the spectrum (2008). Empirical work focusing on mainstream and separatist party rhetoric on self-government would give insight into an important point which is left unspecified here: how does self-government get on the political agenda in the first place? What persuades governments (and potential government parties) to support constitutional reform which reduces their own powers?  Third, the theory is tested by observing the behaviour of separatist parties. The electoral rhetoric and legislative activities of separatists are theorized as being the output of a complex causal process. I argue that separatist leaders start with preferences between a hypothetical and the current political system, plus secondary policy preferences and then consider the opportunities open to them to achieve these objectives over the short and long term. Due to the focus on observable behaviour, I don’t test the early links in this causal chain, meaning we do not have evidence that separatist parties actually did see their strategic position in the way I categorize it, nor that they had the preference ordering I assume.   Chapter 3.2 addresses this question briefly by looking at how the IPP explained the structure of electoral competition to voters, as well as detailing what it thought it could achieve with pivotal status in the British House of Commons. These explanations may have been thought necessary in the 1886 election in particular, as the franchise expansion of the previous year meant there were large numbers of new voters. It would also be useful to examine the extent (if at all) that modern parties explain their electoral and legislative position to voters. A further issue is that the party leadership is unlikely to share all its analysis of its own position with the wider public. So, a further test of the theory would be 193 to look for evidence of internal party debate and planning: either among diaries and personal letters in the case of the IPP or through interviews with BQ and SNP officials and candidates.  A final limitation is that the theory rests on separatist parties’ expectation of voter behaviour. I argue that the leadership focuses on how voters will react to the party’s participation in central institutions, and on how to attract non-separatist voters. The research here does not look at the voters’ own responses to credit-claiming or grievance-based rhetoric, nor to disengagement or participation in parliament. Doing so would help to assess whether the party’s perception of voters in the region is accurate. It would also test the theory’s assumptions about the structure of the electorate in separatist regions: is there a clear division between separatist and non-separatist voters, or are there floating voters who are open to persuasion on this issue?   Interestingly, Sorens’ comparative work on the psychology of separatist voting finds that electoral support for secession is more likely in regions which are affluent relative to the rest of the country (2012) – presumably because voters had more confidence in the viability of the hypothetical new state.  This suggests that rhetoric on the positive case for separation might be particularly effective, alongside anti-system grieving which claimed that the separatist region’s wealth was being siphoned off by the central government (as opposed to claims of relative deprivation). Conversely, Sorens work supports the view that credit-claiming does risk the long-term separatist project. It focuses the attention on the region needing investment from the capital, rather than being able to rely on its own resources.   6.4  Future research The findings presented here show that separatist parties are rational actors whose behaviour can be predicted, while also demonstrating that their motivations and strategic position are distinct from 194 mainstream parties or indeed other types of small parties. This research agenda can be developed by addressing the limitations noted in the previous section and by further theorizing separatists’ legislative behaviour when self-government is off the agenda.  Extending the research to multi-party systems and more permissive electoral systems will test whether the theory still holds when electoral competition is more fluid. An important element of looking at these cases is that small parties (including separatists) are more likely to be pivotal under such circumstances. Although I found only limited evidence that pivotal status affected behaviour, this may be because I was examining systems where majority government is the norm. So, minority governments may see their position as a temporary aberration. In systems where it is rare for one party to win a majority of legislative seats, cross-party bargaining is more common, and, crucially, expected by voters. Two particularly interesting cases to examine would be the Lega Nord and Convergència i Unió, which joined the Italian and Spanish governments respectively. The research questions here would be whether they only did this in periods when there was no reasonable prospect of self-government, and what policy concessions they extracted. Did they try to put self-government on the agenda, or did they focus on secondary policy priorities, institutional reforms in the current system or additional spending for their region?   A second project would test the theory’s assumptions on the behaviour of separatist voters, asking how they actually react to separatist parties’ participation in central legislatures.  Do separatist parties correctly anticipate their voters’ reaction? Such research would use attitudinal data from election surveys to examine separatist voters’ policy preferences and their assessment of separatist party leaders. It would also draw on experiments to assess the cognitive impact of different rhetorical strategies on different types of voter.   195 Finally, it is worth returning to the finding that pivotal status only appears to have an impact when self-government is imminent. So, we see a pivotal party bargaining on the process of devolution, when that is on the agenda. However, in periods when self-government is off the agenda, we do not see separatist parties changing their parliamentary activities. Furthermore, there is a mis-match between legislative position and electoral rhetoric: we have seen that the use of credit-claiming hangs on electoral competition rather than pivotal status. However, we know that separatists’ behaviour in the electoral realm is neither erratic nor random. So, future research should consider how we can systematically explain legislative behaviour in all scenarios.   One possibility is investigating the impact of the separatist party’s caucus size. The 2015 British election offers an intriguing opportunity to study this, as the SNP went from 6 seats at Westminster to 56 (holding all but 3 of the Scottish seats). The speeches of the party’s leaders in the immediate aftermath emphasized that they claimed the exclusive right to speak for Scotland (BBC, 2015).  However, they do not have pivotal status, and the SNP leader has stated that Scottish independence is off the immediate agenda after the failure of the 2014 referendum (BBC, 2014). So, the SNP has had a sudden surge in its electoral appeal and its efficiency in winning seats, with accompanying expectations of what it can achieve for Scotland – but it does not have direct parliamentary leverage. The upcoming parliament will therefore be a window on how separatists deal with a mismatch between electoral and legislative power, and the paradox of electoral success within a political system they want to leave.    196 References Academic works Aldous, Richard. The lion and the unicorn: Gladstone vs Disraeli. London (England): Hutchison. 2006.  Anderson, Cameron D. and Elisabeth Gidengil. Explaining the vote for sub-state nationalist parties: The SNP and the Bloc Québécois compared. Paper presented at the Annual Conference of the Political Studies Association. Aberdeen, 2002.    Axelrod, RM. “The evolutionary approach to norms.” American Political Science Review. Vol 80, No. 4 (Dec. 1986), pp 1095-1111.   Axelrod, RM. The complexity of cooperation: agent based models of competition and collaboration.  Princeton, N.J.: Princeton University Press, 1997.  Axelrod, RM and WD Hamilton. “The evolution of cooperation.” Science, Vol. 211 (March 1981), pp 1390-1396.  Ayres, William R. and Stephen M. Saideman. “Is separatism as contagious as the common cold or as cancer? Testing the international and domestic determinants of secessionism.” Nationalism and Ethnic Politics. Vol. 6, No. 3, (2000), pp 99-113.   Bélanger, Eric and Meguid, Bonnie M. “Issue salience, issue ownership and issue-based vote choice.” Electoral Studies. Vol. 27 (Sep 2008), pp 477-91.  Bennie, Lynn, Jack Brand and James Mitchell. How Scotland votes: Scottish parties and elections. Manchester; New York: Manchester University Press. 1997.   Blidook, Kelly. “Exploring the role of “legislators” in Canada: Do Members of Parliament influence policy?” Journal of Legislative Studies. Vol. 16 No.1 (2010), pp. 32-56.   Bohrnstedt, George W. and David Knoke. Statistics for social data analysis.  Itasca, Illinois: F.E. Peacock Publishers. 1982.  Brancati, Dawn. “Decentralization: fueling the fire or dampening the flames of ethnic conflict and secessionism?” International Organization. Vol. 60, No. 3 (2006), pp 651-685.   Budge, Ian, Hans-Dieter Klingemann, Andrea Volkens, Judith Bara and Eric Tanenbaum. Mapping policy preferences: Estimates for parties, electors, and governments, 1945-1998. Oxford: Oxford University Press. 2001.  Burgess, Michael. Comparative federalism: theory and practice. New York, N.Y.: Routledge. 2006.   Cairney, Paul. The Scottish political system since devolution: from new politics to the new Scottish government. Exeter, U.K.: Imprint Academic. 2011.  197 Cairns, Alan C. “The electoral system and the party system, 1921-1965.” Canadian Journal of Political Science. Vol. 1, No. 1 (1968), pp. 55-80.  Cardinal, Mario. Trans: Ferdinanda Van Gennip with Mark Stout. Breaking point Quebec-Canada: The 1995 referendum.  Montreal: Bayard Canada. 2005.   Carty, R.K., William Cross and Lisa Young. Rebuilding Canadian party politics. Vancouver: UBC Press. 2000.   Clark, Alan. Diaries. London: Phoenix. 1993.  Cox, Gary W. 1987. The efficient secret: the cabinet and the development of political parties in Victorian England. Cambridge; New York: Cambridge University Press. 1987.  Cox, Gary W. Making votes count: strategic co-ordination in the world’s electoral systems. Cambridge: Cambridge University Press. 1997.   Cox, Gary.  “The organization of democratic legislatures.” in Barry Weingast and Donald Wittman, eds. The Oxford Handbook of Political Economy. Oxford: Oxford University Press. 2005.   Craig, F.W.S. British parliamentary election results, 1974-1983. Chichester, West Sussex: Parliamentary Research Services, 1984.  Craig, F.W.S. British parliamentary election results, 1885-1918. Aldershot, Hants. England: Parliamentary Research Services, 1989.   Crick, Bernard. “On conflict and opposition.” in Barker, Rodney (ed). Studies in opposition. London: The Macmillan Press, 1971, pp 38-44.  Cross, William and Young, Lisa. “The contours of political party membership in Canada.” Party Politics. Vol. 10, No. 4 (2004), pp 427-444.   Cunningham, David E. “Veto players and civil war duration.” American Journal of Political Science, Vol. 50, (September, 2006), pp 875-92.  Curtice, John and Ben Seyd. Has devolution worked? : The verdict from policymakers and the public. Manchester U.K.: Manchester University Press, 2009.   Detterbeck, Klaus and Eve Hepburn. “Party politics in multi-level systems” in Erk, Jan and Wilfried Swenden (eds). New Directions in Federalism Studies. Abingdon, Oxon: Routledge, 2010, pp 106-125.  Dewan, Torun and Arthur Spirling. “Strategic opposition and government cohesion in Westminster democracies.” American Political Science Review. Vol. 105, No. 2 (2011), pp 337-358.  Dinas, Elias, Pedro Riera and Nasos Roussias. “Staying in the first league: Parliamentary representation and the electoral success of small parties.” Political Science Research and Methods. (2014), pp 1-18.   198 Docherty, David, C. Mr Smith goes to Ottawa: Life in the House of Commons. Vancouver: University of British Columbia Press, 1997.   Dod’s parliamentary companion guide to the general election, 2005. London: Dod’s Parliamentary Communications. 2005.   Duverger, Maurice. Political parties.  New York: Wiley, 1954.   Ezrow, Lawrence. Linking citizens and parties: How electoral systems matter for political representation. Oxford: Oxford University Press. 2010   Fearon, James D and David D. Laitin. “Violence and the social construction of ethnic identity.” International Organization.  Vol. 54 No. 4 (2000), pp 845-877.   Flanagan, Tom. Waiting for the wave: the Reform Party and the conservative movement. Toronto: Stoddart. 1995.   Fusaro, Anthony. “Two faces of British nationalism: The Scottish National Party and Plaid Cymru Compared.” Polity. Vol 11 No. 3 (1979), pp 362-86.   Galatas, Steven E. Electing the first Parliament: Party competition and voter participation in Scotland. Party Politics, Vol. 10, No. 2 (2004), pp 213-233.   Gallagher Cunningham, Kathleen. “Divide and conquer or divide and concede: How do states respond to internally divided separatists?” American Political Science Review, Vol. 105, No. 2, (May 2011), pp 276-297.   Gidengil, Elisabeth. Dominance and decline: making sense of recent Canadian elections. Toronto: University of Toronto Press. 2012.   Giordano, Benito. “A place called Padania? The Lega Nord and the political representation of northern Italy.” European Urban and Rural Studies. Vol. 6, No. 3 (1999), pp 215-230.    Gold, Thomas W. The Lega Nord and contemporary politics in Italy. New York: Palgrave Macmillan, 2003.   Godbout, Jean-Franҫois and Bjørn Høyland. “Legislative voting in the Canadian Parliament.” Canadian Journal of Political Science. Vol.44, No.2 (2011), pp 367-388.  Greer, Scott L. Nationalism and self-government: the politics of autonomy in Scotland and Catalonia. Albany: State University of New York Press. 2007.   Guibernau, M. “Spain: Catalonia and the Basque Country”. Parliamentary Affairs. Vol.53, No. 1, (2000), pp 55-68.    Gunther, Richard and Diamond, Larry. “Species of political parties: A new typology.” Party Politics, Vol. 9, No. 2, (2003), pp 167-199.  199 Harbom, Lotta, Erik Melander and Peter Wallensteen. “Dyadic dimensions of armed conflict, 1946-2007.” Journal of Peace Research. Vol.45 No. 5 (2008), pp 697-710.  Hattersley, Roy. Fifty years on. A prejudiced history of Britain since the War. London: Abacus. 1997.   Hix, Simon and Noury, Abdul. Government-opposition or left-right? The Institutional determinants of voting in legislatures.  Unpublished working paper, 2014. Available on-line at: http://personal.lse.ac.uk/Hix/Working_Papers/Hix-Noury-GOorLR-12June2014.pdf Retrieved 21 May, 2015.  Henig, Simon and Lewis Baston. The political map of Britain.  London: Politico’s. 2002.   Hepburn, Eve (ed).  New challenges for stateless nationalist and regionalist parties. Abingdon, Oxon: Routledge. 2011.  Horowitz, Donald L. “Patterns of ethnic separatism.” Comparative Studies in Society and History. Vol. 23, No. 2 (1981), pp 165-195.   Horowitz, Donald L. Ethnic groups in conflict.  Berkeley, CA: University of California Press, 1985.  Horowitz, Donald L. “Democracy in divided societies.” Journal of Democracy, Vol. 4 No. 4 (October 1993), pp 18-38.  Horowitz, Donald L. “Constitutional design and the problem of adoption:  Proposals versus processes” in Reynolds, Andrew (ed), The architecture of democracy: Constitutional design, conflict management and democracy. Oxford: Oxford University Press, 2002.   Horowitz, Donald L. “Electoral systems: A primer for decision-makers.” Journal of Democracy. Vol. 14 No. 4 (October 2003), pp 115-27.  Horn, Murray J. The political economy of public administration: institutional choice in the public sector.  Cambridge, England: Cambridge University Press. 1995.  Johnston, Richard. “The electoral system and the party system revisited” in Kernerman, Gerald and Philip Resnick (eds). Insiders and outsiders: Alan Cairns and the reshaping of Canadian citizenship. Vancouver, UBC Press, 2005, pp 51-64.  Jacobs, Alan M. Governing for the long term: Democracy and the politics of investment.  New York: Cambridge University Press. 2011.   Kam, Christopher and Indriđi Indriđason. “The timing of cabinet reshuffles in five Westminster parliamentary systems.” Legislative Studies Quarterly. Vol. 30 No.3 (2005), pp 327-363.  Kam, Christopher. Party discipline and parliamentary politics. Cambridge; New York: Cambridge University Press. 2009.   Kee 1972. The green flag: A story of Irish nationalism. London: Weidenfeld and Nicolson. 1972. 200 Keren, Michael. “Political perfectionism and the “anti-system” party.” Party Politics. Vol. 6, No. 1, (2000), pp 107-116.     Kymlicka, Will. Finding our way: Rethinking ethnocultural relations in Canada. Toronto: Oxford University Press, 1998.  Laitin, David D. “Language conflict and violence: The straw that strengthens the camel’s back.” European Journal of Sociology. Vol. 41, No.1 (2000), pp 97-137.  Landry, Bernard. “Avenir politique et constitutionnel du Quebec: temoignages devant la Commission Belanger-Campeau.” Revue parlementaire canadienne. Vol. 14, No. 1 (1991).  Laver, Michael and Kenneth A. Schepsle. Making and breaking governments: Cabinet and legislatures in parliamentary democracies. New York: Cambridge University Press. 1996.   Levy, Roger. “Finding a place in the world economy. Party strategy and party vote: The regionalization of SNP and Plaid Cymru support, 1979-92.” Political Geography, Vol. 14, No. 3 (1995), pp 295-308.  Lijphart, Arend. Democracy in plural societies. New Haven: Yale University Press, 1977.  Lijphart, Arend. Patterns of democracy: Government forms and performance in thirty-six countries. New Haven, CT: Yale University Press, 2012.  Lloyd, Trevor. “Partition-within-partition: The Irish example.” International Journal, Vol. 53, No. 3 (1998), pp 505-21.   Loewen, Peter John, Royce Koop, Jaime Settle and James H. Fowler. “A natural experiment in proposal power and electoral success.” American Journal of Political Science. Vol.58 No.1 (2014), pp 189-196.  Lynch, Peter. SNP: The history of the Scottish National Party. Welsh Academic Press, 2002.   Lyons, F.S.L. “The Irish Parliamentary Party” in Farrell, Brian (ed.) The Irish parliamentary tradition. Dublin: Gill and Macmillan. 1973.  MacCormick, Neil. “Is there a constitutional path to Scottish independence?” Parliamentary Affairs, Vol. 53, No. 4 (2000), pp 721-36.  Marland, Alex, Thierry Giasson and Jennifer Lees-Marshment (eds). Political marketing in Canada. Vancouver: UBC Press. 2012.  Martin, Lanny W. and Randolph T. Stevenson. “Government formation in parliamentary democracies.” American Journal of Political Science. Vol.45 No.1 (2001), pp 33-50.   McAllister, Laura. Plaid Cymru: The emergence of a political party.  Seren, 2001.   201 McGovern, Clare. Separatist parties in national parliaments: The Scottish National Party and the Bloc Québécois. Paper presented at the annual conference of the American Political Science Association, Toronto, Ontario, September 2009.   McLean, Bob. “Labour in Scotland since 1945: Myth and reality.” in Hassan, Gerry (ed). The Scottish Labour Party: History, institutions and ideas. Edinburgh: Edinburgh University Press, 2004.   Meguid, Bonnie M. “Competition between unequals: The role of mainstream party strategy in niche party success.” American Political Science Review. Vol. 99, No. 3 (2005), pp 347-359.   Meguid, Bonnie M. Party competition between unequals. Cambridge University Press, 2008.   Mitchell, James. Devolution in the UK. Manchester; New York: Manchester University Press. 2009.  Müller, Wolfgang C. and Kaare Strøm. Policy, office or votes? How political parties in Western Europe make hard decisions. Cambridge: Cambridge University Press. 1999.  Nevitte, Neil, André Blais, Elisabeth Gidengil and Richard Nadeau. Unsteady state: The 1997 Canadian federal election. Don Mills, Ontario: Oxford University Press. 1999.  Nordlinger, Eric A. Conflict regulation in divided societies. Centre for International Affairs, Harvard University, 1972.   North, Douglass. Institutions, institutional change and economic performance. Cambridge: Cambridge University Press, 1990.   O’Brien, Joseph V. William O’Brien and the course of Irish politics 1881-1918. London: University of California Press. 1976  Pierson, Paul. “The limits of change: Explaining institutional origins and change.” Governance, Vol. 13, No. 4, (October 2000), pp. 475-499.   Pierson, Paul. Politics in time. Princeton: Princeton University Press, 2004.  Pilkington, Colin. Devolution in Britain today. Manchester: Manchester University Press. 2002.   Rallings, Colin and Michael Thrasher. British parliamentary election results: 1983-1997. Aldershot, Hants: Ashgate. 1999.  Rallings, Colin and Michael Thrasher. British electoral facts 1832-1999. Aldershot, Hants: Parliamentary Research Services, 2000.  Rudolph, Joseph R. and Robert J. Thompson. “Ethnoterritorial movements and the policy process: Accommodating nationalist demands in the developed world,” Comparative Politics, Vol. 17, No. 3 (1985), pp. 291–311  Sartori, Giovanni. “Opposition and control: Problems and prospects.” In Barker, Rodney. Studies in opposition. London: The Macmillan Press, 1971, pp 31-37. 202 Schedler, Andreas. “Anti-political establishment Parties.” Party Politics, Vol. 2, No. 3, (1996), pp 291-312.  Sergeant, John. Give me ten seconds. London: Macmillan. 2001.  Shepsle, Kenneth A. “Studying institutions: Some lessons from the rational choice approach.” Journal of Theoretical Politics, Vol. 1, No. 2, (1989), pp 131-147.   Sorens, Jason. “Globalization, secessionism and autonomy.” in Electoral Politics, Vol. 23, No. 4, (2004), pp 727-52.  Sorens, Jason. “The cross-sectional determinants of secessionism in advanced democracies.” in Comparative Political Studies, Vol. 38, No. 3, (2005), pp 304-26.  Sorens, Jason. Secessionism: Identity, interest and strategy. Montreal: McGill-Queen’s University Press. 2012.  Soroka, Stuart, Erin Penner and Kelly Blidook. “Constituency Influence in Parliament” in Canadian Journal of Political Science, Vol.42, No. 3 (2009), pp 563-591.    Sinpeng, Vipapat. The power of political movement and the collapse of democracy in Thailand. Ph.D. dissertation, University of British Columbia. 2013.  Steiner, Jürg; Bächtiger, André; Spörndli, Markus and Steenbergen, Marco R. Deliberative politics in action: Analysing parliamentary discourse. Cambridge University Press, 2004.   Steinmo, Sven, Kathleen Thelen and Frank Longstreth (Eds). Structuring politics: Historical institutionalism in comparative analysis. Cambridge, England: Cambridge University Press. 1992.   Swenden, Wilfrid and Maddens, Bart (ed). Territorial party politics in Western Europe. Palgrave Macmillan, 2009.   Tambini, Damian. Nationalism in Italian politics: The stories of the Northern League. London: Routledge. 2001   Thatcher, Margaret. The Downing Street years. London, U.K.: HarperCollins. 1993.   Tsebelis, George. Veto players: how political institutions work. Princeton, N.J.: Princeton University Press. 2002.  Valpy, Michael. “ “The crisis”: a narrative.” In Russell, Peter H. and Lorne Sossin (eds). Parliamentary democracy in crisis. Toronto, Ontario: University of Toronto Press. 2009.   Volkens, Andrea, Pola Lehmann, Theres Mattieß, Nicolas Merz, Sven Regel and Annika Werner. The Manifesto Data Collection. Manifesto Project. (MRG/CMP/MARPOR). Berlin: Wissenschaftszentrum Berlin für Sozialforschung (WZB). 2015a.   203 Walker, Brian Mercer. Parliamentary election results in Ireland, 1801-1922. Dublin: Royal Irish Academy. 1978.  Walter, Barbara F. “Building reputation: Why governments fight some separatists but not others.” American Journal of Political Science, Vol. 50 No. 2 (2006), pp 313-30.  Walter, Barbara F. “Information, uncertainty and the decision to secede.” International Organization, Vol. 60 No. 1 (2006), pp 105-135.   Warwick, Paul V. Government survival in parliamentary democracies. Cambridge: Cambridge University Press. 1994.   Warwick, Paul V. Policy horizons and parliamentary government. New York: Palgrave Macmillan. 2006.  Wicks, Elizabeth. Evolution of a constitution: Eight key moments in British constitutional history. Portland, Oregon: Hart Publishers. 2006.  Wood, JR. “Secessionism: A comparative analytical framework” in Canadian Journal of Political Science, Vol. 14, No. 1, (1981), pp 107-134.  Wooldridge, Jeffrey M. Introductory econometrics: A modern approach. Cincinatti, Ohio: Thomson South-Western. 2003  Young, Lori and Bélanger, Éric. “BQ in the House: The nature of sovereigntist representation in the Canadian Parliament.” Nationalism and Ethnic Politics, 14: 4 (2008), pp 487-522.   Young, Robert Andrew. The struggle for Quebec: From referendum to referendum?  Montreal: McGill-Queen’s University Press. 1999.   Government and parliamentary publications  Commission on Scottish Devolution. Serving Scotland better: Scotland and the United Kingdom in the 21st century. Edinburgh: Commission on Scottish Devolution. 2009.  Elections Canada. Official Voting Results. Thirty-Eighth General Election 2004. Table 9: Percentage of valid votes by political affiliation.  2004. Available on-line at: http://www.elections.ca/scripts/OVR2004/default.html Retrieved 21 May, 2015.  Elections Canada. Official Voting Results. Thirty-Ninth General Election 2006. Table 9: Percentage of valid votes by political affiliation.  2009. Available on-line at: http://www.elections.ca/scripts/OVR2006/default.html Retrieved 21 May, 2015.  Elections Canada. Official Voting Results. Fortieth General Election 2008. Table 9: Percentage of valid votes by political affiliation. 2008. Available on-line at: 204 http://www.elections.ca/scripts/OVR2008/default.html Retrieved 21 May, 2015.  The Electoral Commission.  Election 2005: constituencies, candidates and results. 2006. Available on-line at: http://www.electoralcommission.org.uk/publications-and-research/election-reports Retrieved 15 March, 2013.  The Electoral Commission. General election 2010 results. 2010. Available on-line at: http://www.electoralcommission.org.uk/__data/assets/excel_doc/0020/105725/GE2010-constituency-results-website.xls Retrieved 15 March, 2013.  Government Statistical Service. Unemployment statistics from 1881 to the present day. London, U.K.: Her Majesty’s Stationery Office. 1996.  Hansard (debates of the House of Commons of Canada).  Available on-line at: http://www2.parl.gc.ca/housechamberbusiness/ChamberSittings.aspx?View=H&Language=E&Parl=39&Ses=1 Retrieved 21 May, 2015.  Hansard (the official report of the UK Parliament). Available on-line at: http://www.publications.parliament.uk/pa/pahansard.htm Retrieved 21 May, 2015. Hansard (the official report of the UK Parliament). Records prior to 2005 available on-line at: http://hansard.millbanksystems.com/commons/ Retrieved 21 May, 2015.  Marleau, Robert and Camille Montpetit (eds). House of Commons procedure and practice. Ottawa: House of Commons. 2000.  Office for National Statistics. Labour Force Survey, May 2001. London: Her Majesty’s Stationery Office. 2001.  Parliament of Canada. Status of House business, First Session, 35th Parliament. Ottawa: House of Commons. February 2 1996.  Available on-line at: http://www.parl.gc.ca/content/hoc/House/351/status/status.pdf Retrieved on 21 May, 2015.  Parliament of Canada. Status of House business, Second Session, 35th Parliament. Ottawa: House of Commons. April 27, 1997.  Available on-line at: http://www.parl.gc.ca/content/hoc/House/352/status/status.PDF  Retrieved on 21 May, 2015.  Parliament of Canada. Status of House business, Third Session, 37th Parliament. Ottawa: House of Commons. May 23, 2004. Available on-line at: http://www.parl.gc.ca/HousePublications/Publication.aspx?Pub=status&Language=E&Mode=1&Parl=37&Ses=3 205 Retrieved on 21 May, 2015.  Parliament of Canada. Status of House business, First Session, 38th Parliament.  Ottawa: House of Commons. November 29, 2005. Available on-line at: http://www.parl.gc.ca/HousePublications/Publication.aspx?Pub=status&Language=E&Mode=1&Parl=38&Ses=1 Retrieved on 21 May, 2015.  Parliament of Canada. Party standings in the House of Commons. Ottawa: House of Commons. 2013. Available on-line at: http://www2.parl.gc.ca/Parlinfo/lists/PartyStandings.aspx?Menu=HOCRepresentation&Section=03d93c58-f843-49b3-9653-84275c23f3fb Retrieved on 15 March, 2013.  Parliament of Canada. Members of the House of Commons, 1867-date database. Ottawa: House of Commons. 2014. The database is available at: http://www.parl.gc.ca/Parlinfo/Lists/Members.aspx?Language=E&Current=True Retrieved on 21 May, 2015.  Parliament of Canada. Annotated standing orders of the House of Commons. Ottawa: House of Commons. 2005. Available on-line at: http://www.parl.gc.ca/ProceduralReference.aspx?Language=E Retrieved on 21 May, 2015.  Scottish Government. Your Scotland, your voice: A national conversation. Edinburgh, Scotland: Blackwell. 2009.    UK Parliament. Scotland Act, 1998.  London: Her Majesty’s Stationery Office. Available on-line at: http://www.opsi.gov.uk/acts/acts1998/ukpga_19980046_en_1 Retrieved 21 May, 2015.  UK Parliament. MPs at dissolution – 2010. London: UK Parliament. 2014. Available on-line at: http://www.parliament.uk/mps-lords-and-offices/mps/ Retrieved 15 March, 2013.  UK Government. Scotland’s Future in the UK. Command Paper 7738. London: Her Majesty’s Stationery Office. 2009   News stories  BBC Online Archive. 1978: Callaghan accused of running scared. 7 September 1978.  http://news.bbc.co.uk/onthisday/hi/dates/stories/september/7/newsid_2502000/2502781.stm Retrieved 21 May, 2015.  BBC News. SNP outlines plan for referendum. 3 September, 2009. Available at: http://news.bbc.co.uk/2/hi/uk_news/scotland/8233788.stm 206 Retrieved 21 May, 2015.  BBC News. Scottish independence: Cameron “not fussed” over referendum date. 15 May, 2012. Available at: http://www.bbc.co.uk/news/uk-scotland-scotland-politics-18080527 Retrieved 21 May, 2015.  BBC News. Salmond: Independent Scotland “in favour” of EU membership. 20 May 2014.  Available at: http://www.bbc.com/news/uk-scotland-scotland-politics-27482125 Retrieved 21 May, 2015.  BBC News. Scottish referendum: Scotland votes “No” to independence. 19 September 2014.  Available at: http://www.bbc.com/news/uk-scotland-29270441 Retrieved 21 May, 2015.  BBC News. Election 2015: Alec Salmond hails roar of “Scottish lion”. 8 May 2015. Available at: http://www.bbc.com/news/uk-scotland-north-east-orkney-shetland-32641223 Retrieved 21 May, 2015.  The Times. Mr Asquith at the Albert Hall. London, England. Issue 39140, (11 December, 1909), p.8.  Party and candidate documents and speeches  Bloc Québécois. Le Manifeste du Bloc Québécois. 1991. Retrieved from the Poltext project. Centre for the Analysis of Public Policy (CAPP), Université Laval, at: https://www.poltext.org/en/part-1-electronic-political-texts/electronic-manifestos-canada Retrieved, 21 May, 2015.  Bloc Québécois. Le Manifeste. Assemblée de lancement. 1991. Retrieved from the Poltext project. Centre for the Analysis of Public Policy (CAPP), Université Laval, at: https://www.poltext.org/en/part-1-electronic-political-texts/electronic-manifestos-canada Retrieved, 21 May, 2015.  Bloc Québécois. Un nouveau parti pour l’étape decisive. 1993. Retrieved from the Poltext project. Centre for the Analysis of Public Policy (CAPP), Université Laval, at: https://www.poltext.org/en/part-1-electronic-political-texts/electronic-manifestos-canada Retrieved, 21 May, 2015.  Bloc Québécois. Plate-forme électorale. 1997. Retrieved from the Poltext project. Centre for the Analysis of Public Policy (CAPP), Université Laval, at: https://www.poltext.org/en/part-1-electronic-political-texts/electronic-manifestos-canada Retrieved, 21 May, 2015.  207 Bloc Québécois. Un parti propre au Québec. 2004. Retrieved from the Poltext project. Centre for the Analysis of Public Policy (CAPP), Université Laval, at: https://www.poltext.org/en/part-1-electronic-political-texts/electronic-manifestos-canada Retrieved, 21 May, 2015.  Bloc Québécois. Heureusement, ici c’est le Bloc. 2006. Retrieved from the Poltext project. Centre for the Analysis of Public Policy (CAPP), Université Laval, at: https://www.poltext.org/en/part-1-electronic-political-texts/electronic-manifestos-canada Retrieved, 21 May, 2015.  Bloc Québécois. Présent! pour le Québec. 2008. Retrieved from the Poltext project. Centre for the Analysis of Public Policy (CAPP), Université Laval, at: https://www.poltext.org/en/part-1-electronic-political-texts/electronic-manifestos-canada Retrieved, 21 May, 2015.  Corbett, T.L. “To the Electors of East Tyrone.” Belfast News-Letter. 11 July 1895.  Ewart, William. “To the electors of the north division of Belfast.” Belfast News-Letter. 29 June 1886.  Freeman’s Journal. Mr W. O’Brien in South Tyrone. Great meeting at Clogher. 26 June, 1886.  Freeman’s Journal. The St. Stephen’s Green and Harbour divisions. Great popular demonstration. 29 June, 1886.  Freeman’s Journal. South Monaghan election. 10 July, 1886  Freeman’s Journal. Mr Kenny in Tyrone. 10 July, 1886.  Freeman’s Journal. North Down election. Candidature of Mr McNabb. 13 July, 1886.  Freeman’s Journal. The North Monaghan election, great demonstrations at Tullycorbett. 14 July, 1886.  Freeman’s Journal. East Tyrone. Mr Doogan’s candidature. 5 July, 1895.  Freeman’s Journal. The Conventions. East and Mid Tyrone. 8 July, 1895  Freeman’s Journal. County Wexford. Messrs Thomas J. Healy and Peter Ffrench selected. 8 July, 1895.  Freeman’s Journal. Fermanagh. Mr J Jordan and Dr Cassidy selected.  9 July, 1895  Freeman’s Journal. South Down. Selection of Mr Michael McCartan. 11 July, 1895  Freeman’s Journal. Queen’s County (Ossory division). Speech by Dr Tanner – successful canvas.  22 July, 1895  Hamilton, Ernest William. “To the electors of North Tyrone.” Belfast News-Letter. 29 June 1886.  Harland, E.J. “To the electors of North Belfast.” Belfast News-Letter.  11 July 1895. 208 Haslett, James H. “To the parliamentary electors of West Belfast.” Belfast News-Letter. 3 July 1886   Haviland-Burke, E. “To the nationalists of South Dublin.” Freeman’s Journal. 17 July, 1895  Johnson, Hugh H. “To the electors of the Borough of the University of Dublin.” Freeman’s Journal.  2 July, 1886.  Johnston, W.M. “To the electors of South Belfast.” Belfast News-Letter. 9 July 1895.  McCalmont, Hugh. “To the electors of North Antrim.” Belfast News-Letter. 11 July 1895.   McCalmont, James. “To the electors of East Antrim.” Belfast News-Letter. 11 July 1895.   Megaw, Matthew. “To the electors of East Tyrone.” Belfast News-Letter. 25 June 1886.  Mulholland, Henry Lyle. “To the electors of North Derry.” Belfast News-Letter. 25 June 1886.   Plunket Barton, Dunbar. “To the electors of Mid-Armagh.” Belfast News-Letter. 13 July 1895.   Scottish National Party. Assorted election campaign leaflets. February and October 1974. Copies received from University of Aberdeen Special Collections Centre. Reference: LpAK4 SNP 1974.  Scottish National Party. Assorted election campaign leaflets. 1979. Copies received from University of Aberdeen Special Collections Centre. Reference: LfpAK4 Sco  Scottish National Party. Assorted election campaign leaflets. 1983. Copies received from University of Aberdeen Special Collections Centre. Reference: LpAK4 SNP 1983  Scottish National Party. Assorted election campaign leaflets. 2001. Copies received from University of Aberdeen Special Collections Centre. Reference: LpAaK4Ele2001  Scottish National Party. Elect a local champion. Manifesto 2010. Edinburgh, Scotland: SNP. 2010.  Retrieved on-line 10 October, 2013.  The Times. Manifesto to Irish electors.  London, England. Issue 37905, (1 January, 1906), p. 7.  The Times. The crisis. London, England. Issue 39145, (17 December, 1909), p7.  The Times. Mr J Redmond. London, England. Issue 39157, (31 December, 1909), p. 6.   The Times. Nationalist Manifesto. London, England. Issue 39443, (30 November, 1910), p.12.  209 Appendices Appendix A  Case selection Table A. 1 Elections examined in Chapter 3 Party and  election Self-government  on the agenda?139 Separatist party pivotal in previous Parliament? Separatist party’s main competitor140 SNP, October 1974 YES YES  Government party majority at dissolution141: -36  SNP’s seats: 7 Government  Regional seat break-down at last election (Feb 1974):  Labour (Gvt): 40 Cons (Opp): 21 SNP: 7 Lib (Opp): 3 SNP, 1979 NO  YES   Government party majority at dissolution: -10  SNP’s seats: 11 Government  Regional seat break-down at last election (Oct 1974):  Labour (Gvt): 41 Cons (Opp): 16 SNP: 11 Lib (Opp): 3 IPP, 1886 YES YES  Government party majority at dissolution: -216  IPP’s seats: 86 Opposition  Regional seat break-down at last election (1885):  IPP: 84 Cons (Opp): 16 Ind Cons (Opp): 2 Lib (Gvt): 0                                                   139 See Section 3.1.1 for the classification of self-government as being “on” or “of” the agenda. 140 This is measured by the identity of the party which took the plurality of the region’s seats, or if that was the separatist party itself, then the party taking the second highest number. In some cases, the separatist party will not necessarily be the largest party’s main rival – the separatist party may be the third or fourth largest in the region, or may not have existed at the previous election (e.g. the BQ in 1988). The key point is, based on the results of the previous election, which party is most likely to stop the separatist party winning seats? 141 These figures are calculated by subtracting all non-government MPs from the number of government party MPs, excluding the Speaker who does not vote and (in the UK after 1979) Sinn Féin or Independent Republican MPs, as they do not take their seats. The calculations take account of any by-elections, defections or changes in Speaker that occurred during the Parliament as well as seats vacant at dissolution. 210 Party and  election Self-government  on the agenda? Separatist party pivotal in previous Parliament? Separatist party’s main competitor IPP 1895 NO YES  Government party majority at dissolution: -136  IPP’s seats: 81 Opposition  Regional seat break-down at last election (1886): IPP: 80 Unionist/Lib Unionist (Opp): 23 Lib (Gvt): 0 SNP, 2010 YES NO  Government party majority at dissolution: 59  SNP’s seats: 7 Government  Regional seat break-down at last election (2005): Labour (Gvt): 40 Lib Dem (Opp): 11 SNP: 6 Cons (Opp): 1 SNP, 2005 NO NO  Government party majority at dissolution: 166  SNP’s seats: 5 Government  Regional seat break-down at last election (2001): Labour (Gvt): 55 Lib Dem (Opp): 10 SNP: 5 Cons (Opp): 1 BQ, 1993 YES (partially) NO  Government party majority at dissolution: 27  BQ’s seats: 9 Government  Regional seat break-down at last election (1988): PC (Gvt): 63 Lib (Opp): 12 BQ: 0 BQ, 1997 NO NO  Government party majority at dissolution: 62  BQ’s seats: 53 Government  Regional seat break-down at last election (1993): BQ: 54 Lib (Gvt): 19 PC (Opp): 1 Ind: 1   211 Party and  election Self-government  on the agenda? Separatist party pivotal in previous Parliament? Separatist party’s main competitor SNP, February 1974 YES (partially) NO  Government party majority at dissolution: 22  SNP’s seats: 2 Opposition  Regional seat break-down at last election (1970):  Labour (Opp): 44 Cons (Gvt): 23 Lib (Opp): 3 SNP: 1  SNP, 1983 NO NO  Government party majority at dissolution: 37  SNP’s seats: 2 Opposition  Regional seat break-down at last election (1979):  Labour (Opp): 44 Lib (Opp): 3  Cons (Gvt): 22 SNP: 2  Sources: Author’s analysis of data found in: Craig 1984; Craig 1989; Dod’s Parliamentary Communications, 2005; Henig and Baston, 2002; Parliament of Canada, 2014; Rallings and Thrasher, 2000; Electoral Commission (UK), 2006; UK Parliament 2014; Walker, 1978.          212 Table A. 2 Elections examined in Chapter 4 Party and  Election Self-government  on the agenda?142 Separatist party pivotal in previous Parliament? Separatist party’s main competitor143 SNP, 2001 NO NO  Government party majority at dissolution144:178   SNP’s seats: 6 Government  Regional seat break-down at last election (1997):  Labour (Gvt): 56 Lib Dem (Opp): 10 SNP: 6 Cons (Opp): 0 SNP, 1983 NO NO  Government party majority at dissolution: 37  SNP’s seats: 2 Opposition  Regional seat break-down at last election (1979):  Labour (Opp): 44 Lib (Opp): 3  Cons (Gvt): 22 SNP: 2 SNP, 1979 NO  YES   Government party majority at dissolution: -10  SNP’s seats: 11 Government  Regional seat break-down at last election (Oct 1974):  Labour (Gvt): 41 Cons (Opp): 16 SNP: 11 Lib (Opp): 3                                                     142 See Section 4.1 for the classification of self-government as being “on” or “of” the agenda. 143 This is measured by the identity of the party which took the plurality of the region’s seats, or if that was the separatist party itself, then the party taking the second highest number. In some cases, the separatist party will not necessarily be the largest party’s main rival – the separatist party may be the third or fourth largest in the region, or may not have existed at the previous election (e.g. the BQ in 1988). The key point is, based on the results of the previous election, which party is most likely to stop the separatist party winning seats? 144 These figures are calculated by subtracting all non-government MPs from the number of government party MPs, excluding the Speaker who does not vote and (in the UK after 1979) Sinn Féin or Independent Republican MPs, as they do not take their seats. The calculations take account of any by-elections, defections or changes in Speaker that occurred during the Parliament as well as seats vacant at dissolution.   213 Party and election Self-government  on the agenda? Separatist party pivotal in previous Parliament? Separatist party’s main competitor IPP, 1906 NO NO  Government party majority at dissolution: 92  IPP’s seats: 78 Government  Regional seat break-down at last election (1900): IPP: 76 Cons/Unionist (Gvt): 21 Liberal (Opp): 1 Independent Nationalist: 4 IPP, 1909-10 NO NO  Government party majority at dissolution: 99  IPP’s seats: 81 Opposition  Regional seat break-down at last election (1906): IPP: 81 Cons/Unionist (Opp): 17 Liberal (Gvt): 1 Independent Unionist: 3 Independent Nationalist: 1 IPP, Dec 1910 NO  YES   Government party majority at dissolution: -117  IPP’s seats: 71 Government  Regional seat break-down at last election (1909-10): IPP: 71 Cons/Unionist (Opp): 21 Liberal (Gvt): 1 Independent Nationalist: 12 BQ, 2004 NO NO  Government party majority at dissolution: 41  BQ’s seats: 33 Government  Regional seat break-down at last election (2000): BQ: 38 Lib (Gvt): 36 Cons/Alliance (Opp): 1 NDP (Opp): 0 BQ, 2006 NO YES  Government party majority at dissolution: -40  BQ’s seats: 53 Government  Regional seat break-down at last election (2004): BQ: 54 Lib (Gvt): 21 Cons (Opp): 0 NDP (Opp): 0     214 Party and  election Self-government  on the agenda? Separatist party pivotal in previous Parliament? Separatist party’s main competitor   BQ, 2008 NO YES  Government party majority at dissolution: -53  BQ’s seats: 49 Government  Regional seat break-down at last election (2006):  BQ: 51 Cons (Gvt): 10 Lib (Opp): 13 NDP (Opp): 0 Independents: 1  Note: The closeness of the Conservatives and Liberals in Québec in 2006 means that seat count in the previous election is not a clear indictor of the BQ’s competitor in 2008. I therefore examined opinion poll data for the six months prior to the 2008 election by searching Canadian News Stand (a database of Canadian newspapers) for any story in this time period containing the words “election” “Quebec” and “poll”.  I then reviewed the stories for poll results that reported voting intention for Quebec alone (as opposed to the whole of Canada).   These stories consistently showed the BQ and Conservative Party of Canada as being within 1 % point of each other for projected share of the vote in Quebec. The Liberals were either portrayed as being “well behind” the BQ and CPC, or, where the Liberal’s projected share of the vote was given, it was 7-10% points behind the CPC and BQ.   I therefore classify the Conservatives (the incumbent government) as being the Bloc’s primary competitor in 2008. Sources: Author’s analysis of data found in: Craig 1984; Craig 1989; Dod’s Parliamentary Communications, 2005; Elections Canada 2004, 2006, 2008; Henig and Baston, 2002; Rallings and Thrasher, 1999; Parliament of Canada, 2014; Walker, 1978.   215 Table A. 3 Parliamentary sessions examined in Chapter 5 Party and  session Self-government  on the agenda?145 Separatist party pivotal in this Parliament? BQ, 1993-95 YES NO Government party majority146: 61 BQ’s seats: 54 BQ, 1996-97 NO NO Government party majority: 62 BQ’s seats: 53 SNP, 1977 YES YES Government party majority: -1 SNP’s seats: 11 SNP, 1979 NO YES Government party majority: -10 SNP’s seats: 11 BQ, 2003-04 NO NO Government party majority: 41 BQ’s seats: 36 BQ, 2004-05 NO YES Government party majority: -44 BQ’s seats: 54  Sources: Author’s analysis of data found in: Craig 1984; Craig 1989; Parliament of Canada, 2014.                                                            145 See Section 5.1 for the classification of self-government as being “on” or “of” the agenda. 146 Calculated at the beginning of the period under study. These figures are calculated by subtracting all non-government MPs from the number of government party MPs, excluding the Speaker who does not vote. The calculations take account of any by-elections, defections or changes in Speaker that had occurred since the previous election.  216 Appendix B  Coding electoral rhetoric. B.1 Data sources Election manifestos and leaflets can be thought of as a contract between a party and voters, where the party leadership commits itself publicly to a programme of policy over the next legislative period and reports back on what it has achieved over the previous period. Manifestos and leaflets provide a useful measure of overall campaign strategy because they allows parties to explain and justify policy proposals in detail and to draw linkages between their positions on a variety of issues. They are free of the institutional constraints of the legislature (where the government party usually controls the agenda), do not have the space or cost limits of advertisements or social media and are not open to the misinterpretation that may occur in media interviews. A manifesto can therefore be thought of as the party elite’s definitive statement of its position at a particular point in time. Of course, the tone and length of manifestos varies immensely over my party cases: from the IPP’s 14-sentence manifesto in the 1909-1910 campaign to the BQ’s 206-page platform in the 2005-06 campaign. Furthermore, the Bloc’s manifestos are issued in French rather than English.147  However, I am focusing on changes to each party’s rhetoric over time, so these cross-party variations in language, tone and length should not undermine the validity of my findings.   Manifestos for the SNP and BQ were downloaded from the Comparative Manifestos Project database148, the Université Laval’s Electronic Political Texts database149 and the parties’ own websites. I used the French versions of the BQ platforms, translating them into English. The Bloc is primarily a                                                  147 The BQ also issued separate English-language platforms which are much shorter than the French version. In 2004 and 2006 the BQ’s leader wrote separate forewords for each of the English and French language platforms whereas in 2008 and 2011 the English forewords were direct translations of the French forewords. For 2004 and 2006, I used the French versions and translated them into English.  148 Volkens et al, 2015a. 149 The database is part of the Poltext project, at the Centre for the Analysis of Public Policy (CAPP), Université Laval. The Poltext project is funded by a grant from the Fonds québécois de la recherche sur la société et la culture. 217 Francophone party appealing to Francophone voters, so using the French platform captures the message that the Bloc is sending to the majority of its voters. I used Google Translate for the initial translation and then edited the resulting text into grammatical English. Elise Leclerc-Gagné checked the final English translation against the French original for accuracy.   Where manifestos were not available (the SNP campaigns in 1974, 1979 and 1983), I instead used a collection of election pamphlets collected and stored by the University of Aberdeen Library. For the comparisons of the SNP’s overall campaign, I used only the pamphlets which had been issued by the SNP’s central office and which did not make reference to an individual MP. Where I was analysing the rhetoric used in specific constituencies, in order to compare SNP-Conservative contests to SNP-Labour contests, I coded only leaflets which explicitly identified either that constituency or the relevant SNP candidate. I also used SNP pamphlets from 2001 to supplement the 2001 manifesto.  In the early 20th century elections examined in Chapter 4, I use the IPP’s election manifestos that were published verbatim in The Times (London edition). However, the party did not issue an election manifesto in 1886 or 1895. In 1886, the leader Charles Parnell spent the entire campaign in England and Scotland, so did not personally address the Irish voters.150  In 1895, the leader, Justin McCarthy did release an election address to his constituents (published in the local newspaper). However, at that time the IPP was split into two factions and McCarthy was considered the leader in name only (Kee, 1972). So, for these two elections, I searched the archives of the Freeman’s Journal for any speeches or election addresses made by any of the IPP candidates.151  The Journal was published in Dublin and was chiefly concerned with Irish current events. It is described by one historian of the period as                                                  150 See the report in the Freeman’s Journal, 25 June 1886. 151 In 1895, this included all the candidates from both the Parnellite and the Anti-Parnellite factions. Candidates for both elections were identified using Walker, 1978. 218 “respectable” and the only Irish newspaper at the time with “anything like an Irish Nationalist viewpoint” (O’Brien, 1976, 14). However, the British and Irish Press Guide for 1883 classifies the Journal as Liberal rather than Nationalist (see O’Brien, 1976, 14, footnote 2).   This means that we can count on the Journal as a reliable source in two regards. First, as it had a bias towards the Liberal and possibly Nationalist view of Irish politics, it is the newspaper most likely to publish the speeches and addresses of Irish Nationalist candidates. This newspaper therefore gives me the best chance of finding as much material as possible. Second, as it was viewed as an adherent to reasonably high journalistic standards, we can place reliance on the fact that the speeches did occur as reported. The Journal’s reports of election meetings either reproduced candidate’s speeches verbatim, or almost verbatim but transposed into the third person. For each candidate, I coded the election address if there was one, or, if not, the speech he made at his nomination meeting, as this would be the first speech of the campaign and therefore analogous to the manifestos of modern parties.  B.2 Categorization of electoral rhetoric When parties communicate with potential voters during an election campaign, they are making descriptive, attributive and prescriptive claims. Descriptive claims are statements about the way the world is and attributive claims make causal arguments about who or what caused that state of affairs. Parties deploy descriptive and attributive claims in order to convince the voter to do something: they are therefore linked to prescriptive claims, which urge a particular course of action. These prescriptions may be for collective action, in the form of a government policy or an individual action, such as casting a vote for a particular party. I classify descriptive claims as either positive or negative. I also classify attributive and prescriptive claims as either pro-system or anti-system.   219 Descriptive claims: positive or negative? Descriptive claims may focus on a policy, the actions of an individual or group, the government’s overall programme or a socio-economic phenomenon in the country or wider world. They may concern the present, arguing that a policy has had an observable effect, or the future, making claims about the impact of anticipated programmes or policies. A neutral descriptive claim contains no evaluation of what it observes: the policy, action or phenomenon is not framed as either beneficial or harmful to the audience. Given that the goal of electoral campaign material is to persuade readers to take a side in a political contest, I expect such neutral claims to be rare in the data I examine. Most descriptive claims in election material will include an argument that the policy or action in question will be either positive or negative in its effects. For separatist parties, the crucial question is whether they argue that a policy, government or individual action has benefited or harmed their region. So, a claim that a policy benefits the country as a whole, or another region, but harms the separatists’ region would be coded as a negative descriptive claim, and vice versa.  Attributive and prescriptive claims: pro or anti – system? Both attributive and prescriptive claims presuppose that the audience already believes that a positive or a negative state of affairs exists.152  An attributive claim identifies someone or something who is responsible for that state of affairs and a prescriptive claim recommends an appropriate response. Both of these types of claims can be used to either attack or defend the current political and constitutional system.   A separatist party’s attributive claim is pro-system if it gives credit to the separatist party’s position in the central legislature. It could do this by claiming that the party protects the region’s interests and puts                                                  152 This may be because a descriptive claim has been made explicitly in the campaign material, or because the attributive or prescriptive claims are built on what is assumed to be common knowledge among the target audience.    220 pressure on other political actors to send benefits to the region. This claim tacitly supports the constitutional status quo as it is concerned with how well the region’s interests are represented in existing, central institutions. However, this is not simply a distinction between positive and negative claims. An attribution of blame for a negative state of affairs can still be pro-system if it identifies the government of the day as the culprit without proceeding to an argument for self-government. Such a claim is of the kind made by mainstream opposition parties as it implies that the correct response is to vote out the incumbent government, not leave the political system. Conversely, a claim is anti-system if it attributes a negative state of affairs to political institutions, the constitution or all other parties, as it implies that more drastic action is required if the region’s fortunes are to improve.   Campaign rhetoric is also built around prescriptive claims, which state that given that the region has either received benefits or suffered harm, the voter should take a particular political stance. Obviously, in a separatist party’s manifesto, this stance will be a vote for that party, but this exhortation can be framed in either a pro-system or an anti-system way. At the pro-system end of the spectrum, the prescriptive claim would tell the audience that the separatist party should be in the central legislature and therefore the voter should support that party and no other in statewide elections. An anti-system prescriptive argument would build on negative descriptive claims and the attribution of blame to the whole political system in order to build a case that the region should not be in the country at all - therefore urging the audience to support the cause of self-government. Casting a vote for the separatist party is therefore not framed as a means to protect the region’s interests in the central legislature. Rather, it is a statement that the region should be self-governing, and that the central government should offer constitutional reforms which at the very least grant the region more autonomy.   221 So, descriptive claims can be positive or negative. Attributive and prescriptive claims can be pro or anti-system. Figure B.1, below, puts these spectra together in order to categorize a party’s overall campaign strategy. If a party makes descriptive claims which refer to the negative impacts on the region, it is basing its electoral message on grievances: these are the bottom two cells. However, this grieving may be directed at the choices made by the government of the day and the actions of individuals or it may be a fundamental challenge to the political system itself. The former is the same type of rhetoric that any opposition party might use, as its aim is for voters to replace the government or change the government’s behaviour, not seek constitutional change. In the case of separatists, this type of anti-government grieving is characterized by negative descriptive claims, attributions to the government of the day, but pro-system prescriptions (see the bottom left cell of Figure B.1). Voters are told that their region is suffering harm and the solution is to send a separatist party to the central legislature which will be more adept at understanding and representing the region’s interests than another (statewide) party that runs in the region. The bottom right cell represents anti-system grieving: the party’s descriptive claims are that policies, actions or programmes have caused harm to its region and this is attributed to the political system as a whole, or all other political parties. Its prescriptive claims urge readers to vote for the separatist party in order to demonstrate support for self-government.   In both of the top two cells of Figure B.1, the separatist party is using descriptive claims that the region has received or will receive benefits. However, only the top-left cell represents credit-claiming, which combines positive descriptive claims with pro-system prescriptions. This is because the separatist party is arguing that benefits have accrued from its efforts to represent the region in central institutions. The top-right cell is simply the positive case for separatism, as it explains the benefits the region will receive under self-government and therefore exhorts the voter to support separation, making no reference to existing central institutions. 222 B.3 Coding electoral rhetoric Coding was a four–step process. First, I analysed the function of the sentence: was it making an descriptive claim about the world (specifically the separatist party’s own region); an attributive claim about who was responsible for the state of affairs in a descriptive claim, or a prescriptive claim about what should be done in response to the state of affairs described?  Second, I coded the overall tone of each type of claim: were descriptive claims arguing that things were good (or better than elsewhere) or bad (or worse than elsewhere) in their region? Similarly, who did attribution claims focus on? Were they attributed to the political system as a whole, a particular party or government or non-political factors such as the region’s people or natural resources? For prescriptive claims, I was interested in whether any proposed policy was one that could be implemented within the existing constitutional set-up or whether it explicitly required greater self-government or outright independence for the region.  Third, each descriptive claim was considered in the context of the whole foreword or leaflet, to ascertain whether this claim was attributed to any person or phenomenon, and whether it was linked to a policy prescription. For an example of how I coded the linkage between a descriptive claim and an attributive claim, consider first this sentence from the IPP’s 1906 manifesto: “Ireland has lost during the last 20 years 700,000 of her population –one-seventh of the whole  – the only civilized country in the world which has lost population during the last 20 years.” IPP Manifesto 1906, sentence 13.  223  Figure B.1 Categorizing campaign messages   Positive descriptive claims: benefits for the region Pro-system claims: attribute credit to separatists and blame to government; prescription: elect separatists to national institutions Anti-system claims: attribute blame to political system; prescription: self-government Negative descriptive claims: harm to the region Credit claiming.  The positive case for separation.  The negative case for separation:  Grieving               (anti-system).  Grieving  (anti-government).   224 This is a claim about the contemporary de-population of Ireland and is therefore coded as a negative descriptive claim about the present (“the present” being 1906). To assess whether this de-population is being attributed to anyone or anything, I consider this sentence’s relationship to the rest of the manifesto, and find that sentence 13 is the response to a rhetorical question posed in sentence 12: “And what has been the result of this system?” IPP Manifesto 1906, sentence 12.  So, de-population is attributed to “this system” and the use of “this” signifies that the manifesto has already set out in more detail what is meant by “the system” – this is found in the previous paragraph: “No serious attempt has been made to govern Ireland in harmony with the wishes of the people, and even the agents in Ireland of the Unionist Government who spoke of governing Ireland according to Irish wishes have been repudiated and denounced.” IPP Manifesto, 1906, sentence 11. Reading these sentences together, it becomes clear that Irish depopulation is being blamed on the Unionist government specifically and, in general, a political system that governs Ireland without regard to Irish wishes. This combination of a negative descriptive claim and attributive claims focused on both the partisan government of the day and the overall political system of government in Ireland. It is therefore categorized as both anti-government and anti-system grieving (falling into the bottom two cells in Figure B.1).   For an example, of looking for linkage between a descriptive claims and a policy prescription, consider sentences 14-16 from the BQ’s 2004 manifesto: “(14) Currently, more than 190 nations around the world have chosen sovereignty. (15) None of these nations has ever given it up (16) The Québec nation itself has everything to gain by taking the same path.” BQ Manifesto 2004, p.3, sentences 14-16. 225 Sentence 16 is making a descriptive claim about Québec which is positive and future-focused (“it has everything to gain”). The final phrase, “taking the same path” refers back to the policy prescription in sentences 14-15, that of sovereignty. Here, the argument is that positive future benefits are contingent on following this prescription. So, the positive descriptive claim and the anti-system prescriptive claim are linked together to put the positive case for separation - the top right cell in Figure B1.    Fourth, policy prescriptions were coded based on their own content. They were categorized as either substantive (a policy that could be implemented by the central legislature without any change to the region’s governing status) or self-government-focused (directly advocating for greater autonomy for the region). Furthermore, all substantive policy prescriptions were coded as to whether they were linked to self-government: i.e. when the foreword or leaflet states that policy X could only be achieved, or is more likely to be achieved if the region had more autonomy. I also recorded direct reasons given for voting for the separatist party (i.e. prescriptive claims aimed at the individual reader). These claims were then categorized in one of three ways. First, if a vote for separatists was characterized as a means to bring benefits to the region and this is not linked to constitutional change, then it is credit-claiming. If a vote for separatists is framed as a protest against the current government it is anti-government grieving. If, however, voters are urged to vote for the party in order to show support for self-government or to protest against the constitutional status quo then it is anti-system grieving.       226 Appendix C  Additional break-downs of electoral rhetoric data: Chapter 3 cases. Table C. 1 Categorizing campaign messages: IPP, 1886  Pro-system Prescriptive/Attributive Claims Anti-system Prescriptive/ Attributive Claims Total Positive Descriptive Claims Credit -claiming: 17 sentences (5.5% of N) Positive case for separation: 106 sentences (34.3% of N) 123 sentences (39.8% of N) Negative Descriptive Claims Anti-government grieving: 75 sentences (24.3% of N) Anti-system  grieving: 111 sentences (35.9% of N) 186 sentences (60.2% of N)  Total 92 sentences (29.8% of N) 217 sentences (70.2% of N) N = 309 sentences Missing cases (sentences coded as neutral): 4   Table C. 2 Categorizing campaign messages: IPP, 1895.  Pro-system Prescriptive/Attributive Claims Anti-system Prescriptive/ Attributive Claims Total Positive Descriptive Claims Credit -claiming:  40 sentences (37.7 % of N) Positive case for separation:  14 sentences (13.2% of N)  54 sentences (50.9% of N)  Negative Descriptive Claims Anti-government grieving:  39 sentences (36.8% of N) Anti-system  grieving:  13 sentences (12.3% of N) 52 sentences (49.1% of N) Total 79 sentences (74.5% of N) 27 sentences (25.5% of N) N =  106 sentences Missing cases (sentences coded as neutral): 3       227 Table C. 3 Categorizing campaign messages: SNP, February 1974.  Pro-system Prescriptive/Attributive Claims Anti-system Prescriptive/ Attributive Claims Total Positive Descriptive Claims Credit -claiming: 25 sentences (11.3 % of N) Positive case for separation: 120 sentences (54.3% of N) 145 sentences (65.6% of N) Negative Descriptive Claims Anti-government grieving: 13 sentences (5.9% of N) Anti-system  grieving:  63 sentences (28.5% of N) 76 sentences (34.4% of N)  Total  38 sentences (17.2% of N) or 12.5% if inc neutral in N 183 sentences (82.8% of N) N =  221 sentences Missing cases (sentences coded as neutral): 82    Table C. 4 Categorizing campaign messages: SNP, October 1974  Pro-system Prescriptive/Attributive Claims Anti-system Prescriptive/ Attributive Claims Total Positive Descriptive Claims Credit-claiming:  17 sentences (15.5% of N) Positive case for separation: 55 sentences (50% of N) 72 sentences (65.5% of N) Negative Descriptive Claims Anti-government grieving: 10 sentences (9.1% of N) Anti-system  grieving: 28 sentences (25.5% of N) 38 sentences (34.5% of N)  Total   27 sentences (24.5% of N)  83 sentences (75.5% of N) N =  110 sentences Missing cases (sentences coded as neutral): 19        228 Table C. 5 Categorizing campaign messages: SNP, 1979  Pro-system Prescriptive/Attributive Claims Anti-system Prescriptive/ Attributive Claims Total Positive Descriptive Claims Credit -claiming:  137 sentences (31 % of N) Positive case for separation: 97 sentences (22% of N)  234 sentences (53% of N) Negative Descriptive Claims Anti-government grieving: 147 sentences (33.4% of N) Anti-system grieving: 59 sentences (13.4% of N) 206 sentences  (46.8% of N)  Total 284 sentences (64.5% of N)  156 sentences (35.5% of N) N =  440 sentences Missing cases (sentences coded as neutral): 105    Table C. 6 Categorizing campaign messages: SNP, 1983.  Pro-system Prescriptive/Attributive Claims Anti-system Prescriptive/ Attributive Claims Total Positive Descriptive Claims Credit -claiming: 9 sentences ( 8.3% of N) Positive case for separation: 64 sentences (59.3% of N) 73 sentences (67.6% of N) Negative Descriptive Claims Anti-government grieving:  16 sentences (14.8% of N) Anti-system  grieving: 19  sentences (17.6% of N)  sentences 35 (32.4% of N)  Total  25 sentences (23.1% of N)  or 17% if inc neutral in N   83 sentences (76.9% of N) N =  108 sentences Missing cases (sentences coded as neutral): 39   229 Table C. 7 Categorizing campaign messages: SNP 2005  Pro-system Prescriptive/Attributive Claims Anti-system Prescriptive/ Attributive Claims Total Positive Descriptive Claims Credit -claiming:  12 sentences (70.6 % of N) Positive case for separation: 1 sentence (5.9% of N) 13 sentences (76.5% of N) Negative Descriptive Claims Anti-government grieving: 4 sentences (23.5% of N) Anti-system grieving: 0 sentences (0% of N) 4 sentences  (23.5% of N)  Total 16 sentences (94.1% of N)   1 sentence (5.9% of N) N =  17 sentences Missing cases (sentences coded as neutral): 0     Table C. 8 Categorizing campaign messages: SNP 2010  Pro-system Prescriptive/Attributive Claims Anti-system Prescriptive/ Attributive Claims Total Positive Descriptive Claims Credit-claiming: 19  sentences (55.9% of N) Positive case for separation: 8 sentences (23.5% of N) 27  sentences (79.4% of N) Negative Descriptive Claims Anti-government grieving: 5 sentences  (14.7% of N) Anti-system  grieving: 2 sentences (5.9% of N) 7 sentences (20.6% of N)  Total 24 sentences (70.6% of N)  10 sentences (29.4% of N) N =  34 sentences Missing cases (sentences coded as neutral): 2230  Table C. 9 Categorizing campaign messages: BQ 1991 and 1993  Pro-system Prescriptive/Attributive Claims Anti-system Prescriptive/ Attributive Claims Total Positive Descriptive Claims Credit-claiming:  8 sentences (22.9% of N) Positive case for separation: 19 sentences (54.3% of N) 27  sentences (77.1% of N) Negative Descriptive Cl