Partition is the division of an entity into parts. A political partition has often been considered as an objective description: a previously unified territorial entity is divided into two or more parts. Each is demarcated, perhaps with fences, walls, paint or barbed wire, and official posts, where passports may be demanded. Reactions to a political partition are, however, highly subjective. One sees a homeland broken up; another denies it was ever a shared home. That is why the definition of partition that is most pertinent to the field of self-determination defines the action as the externally imposed division of at least one national community’s territory (O'Leary 2007). It is applied, allegedly, to resolve rival national self-determination claims. Partition is perhaps best understood in contrast with secession. Metaphorically secession often ‘unfastens,’ whereas partition ‘tears’. Unfastening unwinds time to a previous territorial order; tearing, by contrast, involves a new or fresh cut, a rip, a gash, a slash. A political partition is a fresh border cut through at least one community’s national homeland, creating at least two separate units under different sovereigns or authorities.

         The usual justification of a partition is that it will reduce or resolve, a national, ethnic or communal conflict. Its opponents, however, protest the freshness, the novelty, the brutality, and the artificiality of dividing a ‘national’ homeland. In Ireland medical metaphors were used–operation, amputation, dismemberment, vivisection and dissection (Connolly 1975). Protestors take the organic unity of the national territory for granted, a relatively modern sentiment. Pre-modern dynasties, by contrast, treated lands as real estate. The partitions of Poland–in 1772, 1793 and 1795– changed terminological history (Mansergh 1997), generating its pejorative associations, and suggesting the new  legitimacy of the nationalist presumption: each nation has a homeland in which it is entitled to govern itself.

         Today unilateral proposals by outsiders to partition a country are now universally judged unlawful: they are against the international norm of respect for ‘territorial integrity.’ Those inclined to propose partition defend it as ‘political triage;’ claiming that the rump and the amputated portions will do better without each other. Proponents and opponents accept that a ‘fresh cut’ is involved: the proponents hope for surgical precision, the opponents deny it is possible. Ireland was ‘freshly cut’ in 1920, after World War 1. Hungary was partitioned in the same year under the treaty of Trianon–its previous borders within the dual monarchy of Austria-Hungary (1867-1918) were not respected. In 1923 the Treaty of Lausanne (1923) partitioned Kurdistan between Turkey and British Iraq, and the Kurdistan foreseen in the Treaty of Sèvres (1920) never materialized. Multiple fresh cuts were made across the empires of the Habsburgs, the Ottomans and the Czars.  Until 1945 the partition of Ireland was widely judged a success by British policymakers–though not by Irish nationalists. After World War II, the partitions of India and Palestine changed the global reputation of partitions, a judgment reinforced by the partition of Cyprus in 1974. 

Distinguishing Partitions and Secessions. While partitions and secessions are often conflated it is historically and analytically necessary to distinguish them. Ireland was partitioned by the British Government in the Government of Ireland Act of 1920, but after subsequent negotiations the Irish Free State was permitted to secede from the Union, and in those same negotiations what became Northern Ireland was permitted to secede from the Irish Free State. The description shows the distinction. The reverse sequence is also possible. A recognized province or region may try to secede, and after the secession is under way a partition of part of a national homeland may occur.

         There is now a Kurdistan Region in Iraq and a Kurdistan Province in Iran, which make secessions possible for these entities, but the establishment of a Kurdistan Region in Turkey or Syria would have to occur before their respective secessions could be even entertained. Therefore a putative partner with a recognized territory secedes; by contrast, an unequal without a recognized territory struggles for liberation. Secessionists have territories; liberationists  must establish theirs. The latter may, however, base their claims on earlier historic jurisdictions, in which case their movement will resemble a secession.

         A partition involves a border adjustment, because there must be a fresh, novel border, whereas a secession involves a border transformation, i.e. the conversion of the previous internal border into a sovereign demarcation. On the understanding advocated here, the breakup of an empire (or of a confederation, or federation, or a union under a common crown) around its existing internal jurisdictions may involve more than one secession, but it is not a partition unless there is at least one fresh cut. 

         Empires and sovereign governments execute partitions. Secession, by contrast, is an action of regions, or provinces, or member-states of a federation or union state, that may be accepted by a political center, or not, in which case we have a contested secession. States have a presumptive right of secession within confederations. By contrast, partition is something governments can execute on a seceding region, against a national liberation movement, or during ‘downsizing’ (O'Leary, Lustick et al. 2001 passim).  The latter is executed by a political center. If that leaves previous provincial borders untouched, there is no partition; it is either decolonization (if there is an organized transfer of authority to a previously unrecognized entity), or dereliction (if there is not).

Types of Partition. Partitions may also be distinguished by whether they divide national or multinational polities; whether they are external or internal; by the agents promoting, supporting and implementing them; and by the political status of the partitioned entities.

         National partitions divide relatively homogeneous homelands, e.g. Germany, Korea and Vietnam at the onset of the Cold War. More debatable examples include the partitions of Mongolia, Kurdistan and Armenia–here previous homogeneity is contested.  National partitions are generally caused by intra-national civil wars accompanied by large-scale interstate wars (or cold wars or foreign interventions) that stabilize the lines of control. These partitions produce schizophrenic entities which claim to be the true embodiment of the nation, and which seek its reunification in their image. In Germany, the capitalist liberal democratic west eventually prevailed; in Vietnam, the communist north–the reverse of what seems likely to occur eventually in Korea. National partitions are initially characterized by mutual non-recognition by the respective regimes, though this may give way to rapprochement and coexistence. Full democratization in the parts of such a divided nation will usually lead to a reunification movement to reverse partition.

         Multinational partitions, by contrast, divide ethnically, religiously, communally or nationally heterogeneous polities (Henderson, Lebow et al. 1974). More nationally homogenous entities may be created, but the deliberate breakup of national or ethnic units within an empire, federation, or a union state is experienced as a national partition for each nationality that has its homeland divided. When the maintenance of heterogeneity within units is the political goal (as in the drawing of the boundaries of some Soviet republics),  and the jurisdictions beneath them (as has been true of the military redesigns of Nigeria’s federation), these cases were national partitions: the goal was to partition as many nations as possible. By contrast, redesigning federations to form internal political borders that correspond with ethno-national homelands or linguistic units may be ‘restorative’ architectures, especially when executed with consensus.

         Internal partitions are driven by one of three strategic goals: control, integration or autonomy. Control includes the deliberate organization of one or more ethno-national groups, and the disorganization and domination of subordinate others: gerrymandering and provincial fragmentation, for example, deliberately dilute the local political concentration of the dominated group(s). Such internal partitions are widespread, e.g. Saddam’s partition of Kurdistan. Integrationist partition, by contrast, carves out heterogeneous units of government with the intention, through ‘mixing,’ of diminishing conflicts between national, ethnic or religious communities. An internal partition may also be organized to promote the autonomy of a particular group that has had no previously recognized jurisdiction.  External partitions, by contrast, involve both the modification of prior homeland jurisdictions, and the transformation of the status of the existing sovereign border. The partitions of Hungary, British India and Cyprus are illustrative.

         Partitions may occur, however, through interactions between outsiders and insiders. Outsiders often want nations divided for military reasons. Where national partitions flow from internal civil wars, the insiders would fight their civil war to the finish until reunification occurred, or peacefully negotiate their reunification, if they could. That is, all insiders, at least initially, regard the partition as temporary. So, unless there is a military stalemate, without outsiders’ actions such partitions will not endure. By contrast, the partitionists of plurinational entities include both outsiders and insiders. The outsiders believe that partition will eliminate (or at least reduce) ethno-national or other identity-based political differences; and they will be supported by at least some insiders who argue the same case. For them, partitions are proposed as long-run resolutions of conflict. Irish nationalists hoped, and still hope, that the partition of Ireland would be temporary. Ulster unionists did not–they could only contemplate its reversal upon Ireland’s return to the Union. Brutish politicians were more ambiguous: some claimed it was a partition of Ireland by the choices of Irish Protestants and Catholics, and could be reversed by their agreement; others hoped that Lloyd George had solved the Irish question.  

         The political status of partitioned entities may be distinguished according to their territories and their peoples. In external partitions, the partitioned territories may be empires or states that have lost wars. In internal partitions, they may be the provinces of union states or federations, entities with equal or at least partnership status with other provinces. Within empires, they may be colonies (held under direct or indirect rule), of inferior status to provinces in the imperial core. The peoples in partitioned territories may be citizens, or conquered or colonial subjects– nomads or hunter-gatherers, were generally not recognized as having any national consciousness about their spaces.

         Some distinguish twentieth-century partitions that occurred during decolonization, e.g. of Ireland, India and Palestine; those that were the product of the Cold War, e.g. of Germany, Korea, Vietnam, China and Taiwan; those executed by a decisive neighborhood power, e.g. Turkey in Cyprus, and, I would add, in Kurdistan; from those that took place in more recent times, namely, those that occurred in the democratization of  the multinational former communist countries. But those who do so, typically conflate partitions and secessions (Schaeffer 1990, Schaeffer 1999).

Implications of Comparisons. The classification just sketched has an important implication. In the later twentieth century, if we leave aside temporary partitions during major wars, there have been far more de-colonizations with colonial borders left intact, and far more secessions, both peaceful and violent, than there have been external partitions (compare with (Sambanis 2000)).  Executed partitions are, of course, much fewer than the number of proposed partitions. Since the end of World War II there have been a very small number of fresh and enduring partitions, including Azerbaijan/Nagorno-Karabakh, Cyprus, Palestine and India. The latter is a huge case, of course, and arguably involved three distinct partitions, of Punjab, Bengal and Kashmir.^[1] The former Yugoslav republic of Bosnia-Herzegovina is a debatable case. Its external borders, within which it seceded from Yugoslavia, were internationally restored, fully intact, when the Dayton Agreement was made in 1995. Internally, however, one of the two entities recognized, namely the Serb Republic, had been created by expulsions and unilateral partition through war, and, in consequence, so was its new ‘partner,’ the Federation of Bosnia-Herzegovina. 

         The small number of recognized external partitions after 1945 reflect two features of the postwar international order. Namely, a general taboo on territorial change by conquest, and on partition by decolonizing powers, after the great bloodbaths between 1947 and 1949 in Palestine and India. Great powers and the UN have generally upheld colonially established borders in Africa and Asia, and have accepted that conquest and prolonged occupation is illegal. It is far easier to have a secession, uncontested or otherwise, recognized (for example, in Croatia, Slovakia, Bosnia-Herzegovina, Macedonia, and the successor states of the USSR) than it is to have a partition recognized (for example, the Turkish Republic of Northern Cyprus (TRNC), and Armenia’s occupation of Azeri territory). Even Kosovo’s secession from Serbia has had far more recognition than the TRNC. External (as opposed to internal) partitions have become taboo. The partition of Ireland was internationally regarded, when noted, as illegitimate in the postcolonial world, ensuring that Irish nationalists had more sympathy among former colonies than Ulster Unionists (Guelke 1988).

Explaining Partition. Most explanations of partitions are indistinguishable from justifications, that is, they explain the conduct of the partitionists as outcomes their ethical and practical beliefs. key feature of these justifications is the pressures that flow from democratization. The latter entails defining the people, the most intractable problem in democratic theory and practice. Externally proposed and imposed partitions of national and of multinational units occur where there is no agreement on who should constitute ‘the people.’

         The world’s borders have been shaped by the great powers. The British had the greatest territorial empire in human history. Irish, Indian and Palestinian nationalists are tempted therefore, for many reasons, to think that partitions are peculiarly British.  British policymakers inclined towards ‘two-nations’ and ‘two irreconcilable religions’ arguments in Ireland, India and Palestine; and they had a previous history of ‘divide and rule’ strategies in their colonies. The British supported and were supported by one minority (Ulster Unionists, Zionist Jews, Indian Muslims and Turkish Cypriots) against the emergent national majority within each colonial entity. The precedent set in Ireland in 1920, and perhaps by some of the frontier adjustments of the League of Nations, subsequently encouraged British imperial elites to think of partition as a viable strategy in Palestine and India.

         One cannot, however, solely hold British imperialists culpable for partitionist enthusiasm; other empires were ‘internally’ partitionist when they seized colonies or merged territories, notably the Soviets. Partitions have been advocated for places that were never inside the British empire: in the Balkans; in the Caucasus; in Africa (e.g., Rwanda and Burundi); and in regions which have left the British empire (e.g., postcolonial Iraq, and Afghanistan). Significantly, however, partitions in these regions have not been (fully) executed, recognized, or (in some cases) even attempted. Cyprus does not neatly fit the hypothesis of British imperial culpability. Its partition occurred after decolonization, and was executed by the Turkish army. Palestine also does not fully fit the hypothesis: the UK’s Foreign and Colonial Offices were divided over the merits of partition, and it was the United Nations, including the US and the USSR, and Abdullah of Transjordan, rather than the British, who provided the final external impetus for partition. A generic and ubiquitous British political disposition to partition cannot be sustained, e.g. they did not partition most of their other multiethnic or their bi-communal colonial territories, such as Sri Lanka, Sudan, Nigeria, and Malaysia. Perhaps by then they had learned that partitions do not work, at least not as intended. However, other decolonizing empires, those run by France, the Netherlands, Spain, Portugal, and the Russians, did not partition their colonial territories on their departures, so there is merit to the idea of British distinctiveness. Proposals were made to partition Algeria, to keep a French enclave, but were rejected. Protecting territorial integrity and promoting linguistic assimilation within their colonial units eventually led the French to be faithful to their Jacobin heritage. The Dutch, the Spaniards and the Portuguese were arguably too enfeebled on exit from their colonies to implement partitions. Both Spain and Portugal were so weak that Morocco and Indonesia, respectively, conquered ‘Spanish Sahara’ and ‘East Timor’– the former conquest has yet to be reversed, the latter has been liberated after Indonesia’s democratization. Remarkably, the Russian Federation resisted the calls of Alexander Solzhenitsyn and others to partition Kazakhstan, despite the proximity of large numbers of Russian and Russian-speaking settlers to the Russian Federation. So far, Russia has not contemplated partition in response to the Chechnya question. Whether this reflects a republican, anti-colonial or Soviet heritage warrants scrutiny. The United States has generally not promoted partitions to resolve ethnic conflicts in its exercise of global hegemony.

         One reason why there may have been a higher incidence of partition under or after British rule is that some meaningful parliamentary government was begun within the empire. Ethno-national mobilization in Ireland occurred against a background of the widening of the franchise. In India, nationalist and communal mobilization also occurred against a widening of the franchise. The Congress party and the Muslim League were beneficiaries of increasing representative and responsible government at provincial level. In Palestine, Jewish settlers were internally democratically organized. In Cyprus, democratic mobilization occurred shortly before decolonization, and competitive pressures among Greek Cypriot politicians made it less likely that the 1960 accommodation with Turkish Cypriots would be maintained. In all these cases democratization encouraged party formation around existing national, ethnic or communal cleavages, thereby making the conciliation of competing demands more difficult, and the formation of ‘a common demos’ problematic. Partition came onto the policy agenda amid rapid ethno-national mobilizations enabled by democratization.

         This hypothesis explains why partition reached the policy agenda but does not explain why it was chosen and then implemented. Blaming ruthless and ambitious political entrepreneurs is a favored theme of historians and political scientists. Applications of this style of thought may be found in many accounts in which partition is seen as the byproduct of a multilayered bargaining game over power and resources between the pro-partition minority, the anti-partition majority and imperial or other third parties. The minority leaders who sought partition in Ireland, Palestine, India and Cyprus are indeed fascinating specimens: Edward Carson and James Craig, Chaim Weizmann and David Ben-Gurion, Muhammad Ali Jinnah, and Rauf Denktash. None began as advocates of partition. Carson, the Dublin-born unionist lawyer, accepted home rule for Northern Ireland and Irish independence with deep reluctance; indeed, he regarded it as a failure. Craig, by contrast, thought a six-county Northern Ireland would be a new impregnable colonial Pale from which to resist Irish nationalism. Both men were ruthless in advocating the abandonment of fellow unionists in counties Monaghan, Cavan and Donegal, the rest of the ancient province of Ulster. A secure majority was more important than incorporating as many unionists as possible. Weizmann and Ben-Gurion were prominent early Zionists, and definers of the prospective boundaries of ‘Eretz Israel.’ They became advocates of the partition of mandate Palestine because they thought it better to have a state than not, and were willing to be ruthless in establishing such a state, whence Ben-Gurion’s interest in ‘transfers.’ Mohammed Ali Jinnah, a Bombay Muslim, did not become an exponent of partition until his sixties. He had been an early, prominent and successful Congressman, and an advocate of secular politics, which he remained. His transformation remains debated. 

         Hardline leaders, solidly endorsed by their most militant and insecure followers, themselves of settler colonial origin, or regarding themselves as of formerly dominant and superior origin, are certainly all parts of the story of partition in Ireland, Palestine, India and Cyprus, but why did such leaders succeed? In Ireland, India and Cyprus, but less so in Palestine, revisionist historiography blames the respective nationalist leaders of the prospective majority communities for placing other priorities ahead of national unity, or for failing to demonstrate inclusive nation-building. Sinn Féin’s leaders (e.g. de Valera, Collins and Griffith) are held responsible for prioritizing sovereignty ahead of the integrity of the national territory, collectively and individually criticized for mishandling the negotiations with Lloyd George’s coalition government,  and for not conciliating unionists.^[2] Congress’s high command is similarly held culpable by historians. Nehru’s underestimation of communalism, Patel’s pandering to Hindu versions of it, and Gandhi’s pervasively Hindu discourses are taken to task, to debunk Congress’s secular self-representation (Anderson 2012). Congress’s leaders placed speedy independence and a strong central government ahead of accommodating the identities, interests and ideas represented in the Muslim League, which it underestimated. Its failure to ensure that the Congress-run provinces accommodated the League after 1937 is much emphasized – it undermined Muslim support for an all-India state. In Cyprus, Archbishop Makarios is held culpable for seeking to unwind the generous settlement reached with Turkish Cypriots, though it is recognized that he risked being outflanked by ultra- and pro-enosis nationalists (as confirmed in the coup that toppled him). (Little revisionist literature, by contrast, criticizes Palestinian leaders for insufficiently accommodating the interests of Zionists.)

         These arguments are reminders of the flaws of the respective nationalist leaders and movements, but they go too far in emphasizing their freedom of choice. These leaders had constituencies. Irish and Indian independence had been long sought and blocked by British imperialists. Irish, Indian and Cypriot nationalisms were all formally civic, and not devoid of initiatives to compromise with their respective minorities. The first two showed willingness to compromise in key negotiations over future institutions, a second chamber,  minority rights, and  a provincial parliament for Ulster within Ireland, and  on a loose federation for India. The Greek Cypriots agreed generous consociational terms for Turkish Cypriots in 1960.

         What responsibilities for partitions lie with mass public sentiments and activities? There is a view that partition is driven by irreconcilable collective identity differences, emanating from long––or recently–established hatreds, inflamed by religious differences. Its corollary is that democratization, decolonization, and the prospect of a new political order, after imperial or dictatorial rule, bring such passions to the fore. It will not do simply to dismiss these theses, as is now habitual among political scientists. To take them seriously is not to commit the sin of primordialism. Social constructionists insist that identities are always flexible, multiple, open, fluid, unpredictable, and not driven by inherited traditions, and tend to deny that ethno-national conflicts are ever rooted in ancient hatreds or less ancient history. But it is usually the case, before  serious ethno-national conflict erupts, that politicians, paramilitaries and others insist on the factuality of the historic maltreatment of their peoples, or credibly warn of future insecurities. These claims and fears must have some resonance with the targeted publics if they are to have results. They cannot be conjured up  ex nihilo by entrepreneurs.  Ethno-national grievances and religious communalism had previous histories in Ireland and India, long preceding democratization. Zionism, secular or otherwise, was a response to Jewish collective grievances, mostly (though not only) at the hands of European rulers. Cypriots were divided by language, religion, and political identification. It is foolish to deny histories of international, interethnic and cross-religious cooperation within places that subsequently were partitioned, but it also makes no sense to deny that collective identities and sentiments, and their expressions in hostile and stereotypical forms, provided necessary and fertile grounds for political mobilizations and counter-mobilizations.

         Whatever the rigidity or longevity or degree of antagonism among collective identities, they do not suffice to explain partitions–though they account for the motivations of the partitionist. In the cases just discussed, partition was chosen by leaders who believed in irreconcilable differences between local parties, and who determined what they hoped would be a final settlement (in conditions of democratization and/or decolonization). They did not allow negotiations between the local agents to be decisive, or claimed they would never produce workable compromises. They implemented such partitions, during or after wars, when downsizing, or decolonizing. 

Justifications of partition. The most powerful arguments to justify partition may be labeled respectively ‘historicist,’ ‘last resort,’ ‘net benefit,’ ‘better tomorrow’ and ‘realist rigor.’ Historicists assume that history is evolving in a given direction, and conclude that it should be given a nudge (Popper 1976 (1957)). In national and ethnic conflicts they insist that once conflicts pass a certain threshold that they will end in partition. They detect such tendencies in residential, educational and employment segregation, or in the formation of nationalist, ethnic or communal parties. Historicist beliefs may shape policy because they are seen as realistic: partition is inevitable; facts have been established ‘on the ground;’ the emergent process should be speeded up to reduce the pain. There is, however, no social science law that all segregation – voluntary or forced – leads inevitably to the breakup of states. Not only has no one identified clear thresholds of violence (absolute or proportional to population) beyond which partition (or separation more broadly) becomes inevitable, many recent conflict-settlements show that there can be peace without partition.

         The last resort argument acknowledges that alternative strategies exist to manage or resolve national, ethnic or communal conflicts, e.g. federalism, consociation, arbitration, or integration. It accepts that these alternatives should be attempted before partition is considered. But if these options fail, so the discourse goes, partition should be chosen to avoid genocide or large-scale ethnic expulsions. Exponents of this argument often invoke the ‘security dilemma.’ In conditions of emergent anarchy, e.g. when an empire or a regime is collapsing, the relations among ethnic groups becomes akin to that of individuals in a Hobbesian state of nature. One distrustful community will seek to enhance its security by preemptive violence, which will enhance the insecurity of the others, creating a vicious and escalating cycle. Ethnic groups with strong and durable identities will mobilize for war, and attack ethnic islands of the other community, or protect their own by expelling others. Partition is justified, it is said, because it ends the imperatives to cleanse and rescue, and renders war unnecessary to achieve mutual security. Some consider partition an appropriate policy choice simply where there are ethnically intermixed populations, because they are capable of sustained pogroms, massacres, expulsions and genocide. This argument, of course, licenses too many partitions: after all, of which groups could it be said that they are incapable of massacres?

         The net benefit argument suggests instead that partition should be chosen when, on balance, it offers a better prospect of conflict reduction, i.e. it should be considered a desirable preventive strategy. This argument was maintained in the last years of British imperial rule by the leading politicians of minorities who opposed independence within existing colonial borders. Partition was justified to prevent a loss of freedom. Exponents did not maintain that genocide and ethnic expulsions were going to be carried out by Irish, Indian or Palestinian nationalists. Rather they feared Roman Catholic, Hindu or Muslim dominance. 

         In the better tomorrow argument, partitionists maintain that after the deed is done there will be a reduction in violence and conflict recurrence. The new and more homogenized polities will have better prospects of stable democratization, of political development, and of better relations with each other. After the trauma is over, the former partners will conduct themselves better, because their interests will not interfere so intimately with one another’s identity, pride and emotions. This argument rests on questionable counterfactual assumptions, namely, that without partition there will be significantly more conflict and conflict-recurrence; and that more heterogeneous polities have poorer prospects of democratization, political development and inter-group relations.

         The tough-minded maintain realist rigor. They claims that any difficulties with partition flow from irresolution—a thoroughgoing revision of borders, which fully separates the relevant antagonistic communities, is what is required. Good fences make good neighbors; bad fences provoke disputes. Policy makers must therefore devise borders—and provide incentives for controlled population movements—to create sufficient homogeneity, so that the incentives for national, ethnic, religious and communal violence are radically reduced. This counsel means: if the initial cut is poor, then make it better.

          These arguments for partition are and have been political and moral. They are not simple apologias, though they may provide cover for more contemptible motives. These arguments are not obviously racist, sectarian, or civilizationist. They present partitionism in its best light. These arguments are only partially testable, however (Sambanis 2000, Chapman and Roeder 2007, Johnson 2008, Sambanis and Schullhofer-Wohl 2009). One accepts historicist philosophies or approaches, or one does not. The leap from demographic trends to assumptions of future political behavior by pro-partitionists is not scientific; the same trends may be compatible with a range of political relationships – from genocide to federal or consociational coexistence. The ‘realistic rigor’ thesis is not testable, because confronted by the evidence of catastrophe, partitionists will always claim that the tragedies lie in the imperfection of the attempted project not the idea itself. An interesting literature now tests partitionist claims in statistical studies, but it is, perhaps unsurprisingly, marred by deep disagreement on the coding of partitions (starting with Sambanis 2000; see comments in O’Leary 2007.  Partitionists implicitly predict either a linear or an exponential relationship between the degree of national and ethnic heterogeneity of a place and the security dilemmas that provoke violence (which can be shown to be false). Ultimately their arguments suggest it is foolish to insist on maintaining unviable multinational polities.

The modalities. How should partitions be performed? Principled partitionists are either proceduralists or paternalists. Proceduralists favor justice and agreement, while paternalists favor imposition in others’ interests–they put order before justice. Proceduralists advocate consultation with the ‘affected parties,’ and try to establish rules to which reasonable partitions should conform. They see roles for commissions, judges and technical experts.

            Honest proceduralists reject proposals that do not meet fairness and feasibility requirements. The requirements of a fair partition have been specified: when it is negotiated by all the affected groups rather than imposed; when it involves a fair division of land and resources; and when it results in homogeneous, or at least substantially less plural, independent states (Lijphart 1984).^[3] The major difficulty with this reasonable conception is the sheer unlikelihood of the first requirement: non-imposition. The affected parties—politicians and their publics—are not likely to agree unanimously, and even if representative politicians did concur, it is unlikely that the adversely affected people will agree, even if offered significant compensation. Partitions involving the movement of people or of their sovereign territory just do not proceed with technical agreement and political consensus.

            These criteria nevertheless offer good benchmarks against which to evaluate the fairness of partitions of bi-national or multinational polities. The Radcliffe ‘Award’ in Bengal in 1947 is unusual because it almost perfectly met the second and third criteria. West Bengal, an area of 28,000 square miles, was to contain a population estimated at 21.19 million people, of whom 29 per cent were Muslims. East Bengal, to become East Pakistan, an area of 49,000 square miles, contained a population of 39.11 million, of whom 29.1 per cent were Hindus. West Bengal was to get 36.6 per cent of the land to accommodate some 35.1 per cent of the Bengal population, while East Bengal was to get 63.6  per cent of the land to accommodate 64.8 per cent of the population. The ratio of majority to minority populations was almost identical, and the resulting entities more homogenous than their predecessor, partitioning a polity with a Muslim-Hindu ratio of 56:44 into two with 70:30 majority: minority ratios. We might equally conclude, however, that Radcliffe created two large Northern Irelands out of Bengal, and few regard the bloody  partition of British India as a success story.

            Others have argued for rules to govern ‘transfers’. Borders should be drawn to leave as few people as possible on the ‘wrong’ state; each individual may emigrate to the ‘right’ state; and each state should be entitled to evict members of the other group. But why should an equal number of ‘wrongly’ placed people be regarded as a fair outcome, as opposed to an equal ratio of ‘wrongly’ to ‘rightly’ placed people? Surely fairness should include proportionality, not just absolute numbers? Such transfer rules, moreover, not only license ethnic expulsions, but also may incentivize them.

            Paternalists, by contrast, assume that the local communities are incapable of agreement. They propose that a sufficiently powerful outsider should determine a durable partition, and reduce conflict credibly—and quickly. Addressing security imperatives is more important than meeting participation requirements or considerations that might flow from abstract social justice.  ‘Better rough justice than none’ is the outlook. It was the outlook of Lloyd George, Winston Churchill and leading lights in the Liberal and Liberal-Conservative coalitions that governed the United Kingdom between 1911 and 1924.

 

The Counter-arguments of Anti-partitionists. Nationalists reject the rupturing of their national territories; multi-nationalists reject the historicist assumptions of homogenizers, and their negative assessments of the prospects for multi-group coexistence. They share, however, common appraisals of how partitions are perverse, of how they jeopardize existing relationships, and of the impossibility of achieving fair partitions.

            The rupturing of national unity is a claim often made against partitions. Ireland, it is rightly said, had been brought ‘whole’ into the union; it is then claimed that it should have been allowed to leave ‘whole.’ In the twentieth century partitions were rejected by most of the affected majority nationalists whose national homelands were freshly cut. They were rejected as undemocratic because they were imposed against the expressed preferences of the relevant majorities in their national territories. They complained that partition was proposed in the interests of privileged minorities. It was especially brutal in its impact on those who became new ‘border communities.’ The Sikhs of the Punjab suffered horribly–and inflicted horrible sufferings themselves. Irish nationalists of south Armagh, Tyrone, Fermanagh, Derry city and Newry suffered less spectacularly, as did their co-nationals in the border counties of the South, in Donegal, Leitrim,  Cavan, Monaghan and Louth, but they undoubtedly suffered.

            Democratic pluralists argue that power-sharing arrangements must be properly exhausted before partition is considered. If there were two nations in Ireland, India, Palestine and Cyprus, or three in Bosnia Herzegovina, no automatic case for partition followed. Home rule within home rule, for example, was possible in Ireland. Power-sharing within Ireland or power-sharing within Northern Ireland was insufficiently explored. The undesirability, infeasibility or insecurity of bi-national, federal, consociational or confederal arrangements must be demonstrated, not assumed, say democratic pluralists. In the British imperial cases, most of the relevant minorities—Ulster unionists, the Muslim League, and Zionists—were deeply unwilling to propose or experiment with such formulae. Their veto of alternative formulae, backed by force, was rendered more effective by the declarations of the imperial power that they would not coerce the relevant minorities. That meant that the hardliners were reinforced, and that the minorities inside these minorities would be coerced. That partition was ‘a last resort,’ or a regrettable choice ‘when all else had failed,’ rings hollow in light of the facts.

            International procedures, including World Court jurisprudence, have addressed some border disputes between states. Typically, however, these arise from ambiguities in historic treaties or legislative documents, or, involve maritime jurisdictions, or are occasioned by natural geographical changes in terrain and river beds. Legal procedures are not, however, appropriate for what is at stake in political partitions. From 1945 until 2009 only two disputes where homelands were arguably at stake, both involving marginal islands, have been settled by the International Court of Justice, one being the Minquiers and Ecrehos Islands located between the English-speaking Channel Islands and French Normandy. At The Hague, the  Permanent Court of Arbitration’s determination of the borders of Abyei, at issue between Sudan and South Sudan,  has not, sadly,  resolved the status of that entity, though it has delimited the focus of dispute.

            No impartial procedure, however, is likely to work amid mass ethno-nationalist politics. Threatening partition encourages preemptive action, especially ethnic expulsions, to establish ‘facts on the ground.’ These repercussions are more likely than bringing the disputing parties to their senses. Pro-partitionists get the causality wrong: partitions generate major security dilemmas, rather than vice versa.

            Anti-partitionists observe that boundary commissions leave pivotality in the key bodies with the relevant big power. The Irish Boundary Commission of 1924-5 and the 1947 Radcliffe Commissions in Punjab and Bengal, had British appointees in the chair. Six related difficulties for such commission chairs can be identified, 

1. Around what around should new boundaries be drawn?In Ireland the question was whether the units should be provinces, counties, parliamentary constituencies or poor-law jurisdictions. Each implied different answers regarding the allocation of places and peoples.
2. Should there be subunit opt-outs? If a majority in unit A wants to be with one state, but B, a concentrated minority within A, wants to go with another, may B opt out?
3. How should units’ preferences be determined?If there is agreement on the units of determination, then how should the new boundary-making respect popular preferences? Through local plebiscites, aggregating across units, or through assumptions about preferences through knowledge of ascriptive identities as recorded in census data (that may be unreliable)? If plebiscites, what rule should be adopted for determining whether a given unit goes to one jurisdiction or another: a simple majority, an absolute majority of registered voters, or a weighted majority? If working from census data, who should count: adults, or adults and children?
4. In determining the allocation of territory around a boundary line should local popular preferences be considered just one criterion to be balanced among others? And if there are other criteria to be considered which other criteria should matter? Contiguity; preserving a cultural heartland; retaining a unit within an economic region or geographical watershed; keeping key built infrastructure within one jurisdiction; or ensuring militarily secure borders? How should these criteria be weighed against popular preferences?
5. If such other criteriaare to be considered in designing new borders, should local popular preferences be subordinated to these other considerations, and, if so, which ones–and who should make that determination?
6. Should there be a ratification process to endorse the work of such commissions, and should there be provisions to enable the subsequent revision of their proposals?

            It is, of course, difficult to imagine impartiality in the appointment and management of a boundary commission.

            Anti-partitionists additionally maintain that partitions encourage ethnic or sectarian expulsions; trigger partially chaotic breakdowns in order, leading to flight, opportunist killing, rapes, and looting; to more violence than was occurring before them; have domino effects; contribute to post-partition wars, and insecurities; and set precedents that lead to demands for repartitions. The case, in short, is that partitions are perverse. They achieve the exact opposite of their goals.  In raw numbers of dead and forcibly displaced, the critics are correct across the cases of India, Palestine, and Cyprus. The partition of India was accompanied by a death toll, variously credibly estimated at between 200,000 and 2 million. Involuntary and expelled cross-border refugees and displaced persons may have approached 15 million.  The partition of Palestine and the war that accompanied Israel’s declaration of independence led to the deaths of approximately 6,000 Israeli Jews, and over 10,000 Arabs, the expulsion and flight of over 750,000 Palestinians, and, as a byproduct, over half a million Jews fled or were expelled from surrounding Arab states. In the Turkish invasion and partition of Cyprus 6,000 Greek Cypriots were killed and 2,000 reported missing, and some 1,500 Turks and Turkish Cypriots killed. After the partition more than 10,000 Greek Cypriots were pressurized into leaving Northern Cyprus, on top of the nearly 160,000 who had fled before the Turkish army. The partition of Ireland was accompanied by the least violence among the twentieth-century partitions. But the violence during and after partition was nevertheless much higher than the death toll before partition, both in the North and the South. Thousands of Catholics were expelled from their jobs and their homes in Belfast and the Lagan valley; there were pogroms in Belfast against Catholics;  both Protestant and Catholic civilians suffered atrocities; and thousands of Protestants emigrated from the Irish Free State.

            Partitions are especially perverse, when they have domino effects—triggering post-partition wars. The Arab-Israeli wars of 1956, 1967, and 1973, and the Israeli-Lebanese wars, prove that the partition of Palestine did not end conflict. India and Pakistan have fought three wars, in 1948, 1965 and 1971.

            Partitions, critics observe, generally do not produce the more homogeneous states envisaged by their advocates. The case literature here is indecisive, however. Israel was more Jewish, and the West Bank and Gaza, more Muslim and Christian, than pre-partition Palestine, but that is partly because the boundary was determined by war and expulsions, not the UN partition plan. Post-partition Israel retained a significant Palestinian Arab minority, and soon had waves of new Jewish refugees of diverse ethnic formation. The units of post-partition Cyprus are very ethnically, linguistically and religiously homogenized by comparison with pre-1974 Cyprus. Post-partition India and Pakistan are vast, populous and multiethnic, and they remain multi-religious. West Pakistan experienced a fresh infusion of linguistically differentiated refugees.  Northern Ireland was created with a unionist and cultural Protestant to nationalist and cultural Catholic ratio of 67:33, which subsequently shifted to 60:40. It  may  slowly move past a current  ratio of 55:45 toward parity. The Irish Free State, however, was more Catholic, Irish and nationalist than pre-partition Ireland, and that Northern Ireland was more Protestant, British and unionist than the nine counties of historic Ulster. But Northern Ireland was not more homogeneous than pre-partition Ireland.

            Critics of partition are especially effective when they argue that partition alone is unlikely to generate the desired or predicted homogenization. Accordingly ‘realists’ push to the surface what others leave as a tacit assumption: the necessity of expulsions to homogenize post-partition entities. Governments that want homogenization  must pursue ‘transfers,’ and are driven to condone or organize expulsions. Post-partition governments may pursue policies that encourage potentially or actively disloyal minorities to emigrate, or encourage the immigration of the ‘right’ people.

            Critics observe that partitions generate new security crises of an inter-governmental form, and cause significant economic disruption–and not just because of communal conflict and warfare, and sudden flows of refugees. Partitions disturb established monetary and exchange networks, increase transactions costs, enhance the likelihood of protectionism, and provide incentives for smuggling and other border-related criminality. They lead to the depreciation of significant capital investments in transport. Roads, railways and canals, ports and airports, terminate original functions, or are damaged. Severe economic losses may flow from failures to cooperate in agriculture, water management, natural resource extraction, and energy production and distribution. The new post-partition entities retain common functional and infrastructural interests,  flowing from their contiguity and their shared pasts, so if they are rational they often end up, ironically, considering post-partition cross-border functional cooperation.

            Of the cases considered here Northern Ireland was not an economic success story before 1998; post-partition Pakistan is acknowledged as a developmental disaster; the story of post-partition Palestine is known to the world; and the unrecognized Turkish Republic of Northern Cyprus has an unenviable reputation. There is a pattern across these cases: one entity (Ireland, India, Israel, and Greek Cyprus) has done better than the other (Northern Ireland, Palestine, Pakistan, Northern Cyprus), though Ireland did not do better economically than Northern Ireland until the late 1980s. Partitionist triage has therefore certainly not been equally good for all.

            Lastly, partition may worsen the ‘compactness’ of the post-partition entities by contrast with their precursors. ‘Compactness’ is the physical solidity of a state—once widely believed to have implications for its military security, it still affects popular assumptions about the right shape of a state, however much academicians reject the thesis of ‘natural boundaries.’

---

^[1] The cases of Abkhazia and South Ossetia are contested secessions, enforced by Russian power over Georgia. They were previous entities in Georgia, though the Georgians treat these secessions as partitions of their homeland. Crimea’s reunification with Russia  is coded as a secession by some, a partition of Ukraine by others: it arguably has elements of both secessions and annexation (because Crime had previously recognized autonomy within Ukraine). 

^[2] For a defense of Ireland’s negotiators in 1921 see Lee: 1998 29-30 Lee, J. J. (1998). The Challenge of a Collins Biography. Michael Collins and the Making of the Irish State. G. Doherty and D. Keogh. Cork, Mercier Press: 19-38.     who observes that the critics forget that Britain carried more powerful guns to the table, and that Sinn Féin achieved far more from the negotiations than Redmond had.

^[3] The journal mistyped the title of this article, which should have read ‘Time and The Politics of Accommodation...’