Thank you for this very useful intervention.

Introducing Lijphart into the discussion is important to giving conceptual clarity to what we are talking about. I have all four elements in mind when thinking about consociationalism, a judicious mixture of which would need expression in a future constitution and political system. For present purposes, however, I have marginally the 2nd (see below), and principally the 3rd and 4th elements in mind. The design of a future central second chamber has to be in the context of the asymmetric autonomy that would have to be negotiated with the Northeast (4th element). Sumane’s argument, in my view, is about the principles governing how we give effect to the 3rd element in the primary institution of power-sharing at the centre – the Senate.

I might add that I have certain strong reservations about the way the 1st and the 2nd element have been given constitutional expression in some systems (e.g. Lebanon). Grand coalitions, including proportionate representation in the executive, are I think more appropriately the subject of political conventions rather than constitutional regulation. On the other hand, the proportionality principle is valuable as a constitutional postulate – even a foundational principle – that imbues the entire culture of government. Here I am thinking about how the ECJ uses proportionality as a principle of the acquis communautaire; the ECtHR in respect of derogations; the German Bundesverfassungsgericht’s jurisprudence not only as part of the Grundgesetz, but also as a fundamental principle of administrative law in the judicial review of executive action; or how the French Conseil Constitutionnel and the Conseil d’Etat use Les Principes Generaux du Droit. Thus proportionality is generally a good idea as a principle of governance, and in our case, would have the contingent effect of promoting inter-community equity and justice (as opposed to the consociational approach, which posits proportionality of representation as the central organising principle of the constitutional order itself.)

In respect of the Belgian system, I would add to the federal government and the Concertation Committee, the Constitutional Court as an important shared rule institution.

Moreover, you also raise a very important point in the design of federal systems here – the question about the number of units. Experienced constitution-makers such as Nicholas ‘Fink’ Haysom and Yash Ghai often make the empirical argument (there being no theoretical one) that the ideal number of units in a federal-type system is six to eight. This number usually provides the matrix for stability and reinforcement of the constitutional structure as a whole in a system that necesarily envisages centripetal and centrifugal tensions (not to mention those between regions themselves) as an institutional feature of the system. Generally, a smaller number of units, especially when conceived as you say as a scheme of (conflicting) identity, generates tensions that the federal constitutional order cannot withstand. Czeckoslovakia is often cited as an example of why two unit federations do not work, although I do not think that the best example. More relevant to the point I make is Belgium (whether or not secession happens), and arguably, the possibiliy of the future break-up of Iraq into three (where there are legally 18 governorates but the constitution allows two or more governorates to merge into extraordinarily powerful regions – thus constitutionalising what is politically a de facto three unit federation between the Kurds, Shi’as and Sunnis). Iraq is also pertinent as a cautionary tale re shared rule – the federal second chamber is left to be set up by statute at a future date!

I am in full agreement with you on the potential benefits of federal-type arrangements on conflict unrelated matters such as economic growth and development. Indeed, it has been my personal experience that the more you go to the periphery (i.e. away from the Western Province), the more receptive people are to localising decision-making. This should almost be obvious, but the argument is not often made.

On the other hand, I do not find problematic the central second chamber understood as an ‘elitist institution’. Democratic elitism has a central role in leadership especially in a the first fragile post-conlict phase and can (although there is no guarantee in fact) help consolidate the new constitutional order. I believe it is also in assonance with the culture of party politics in what is a deeply politicised society.

Finally, you seem to be making a similar point as I was with my initial second critique of Sumane’s proposal: namely, that asymmetrical representation in the Senate may be necessary in addition to asymmetrical regional autonomy for the Northeast. Sumane has responded to me as above, and it will interesting to see what he says to you.

The functions (what you call ‘powers’) of the central second chamber – i.e., what its supposed to do – I agree with you, are central to this discussion. Critical in this respect is its role in constitutional amendment (especially the autonomy convenant), and as the forum in which provision can be made for a Belgian-style ‘alarm bell’ or early warning system, although I am not sure whether this should extend to a veto (although a strong argument can be made that as a minority protection guarantee, the veto is the lynchpin of the device). I am also a strong supporter of the Spanish model as useful for Sri Lanka. That is, at the initial post-conflict constitution-making stage, we focus on the Northeast (including the question of Muslims), and at which stage there is no expressed desire in other parts of the country for autonomy. But the constitution will contain the procedure and the principles through which future regions can be negotiated. This extends also to the possible renegotiation of the original autonomy arrangements between the centre, the Northeast and other areas. in this type of schema, the second chamber is the critical arena, at the political level, for the initiation and negotiation for amendment of the constitution. This is also why I would support the Bundesrat model where the members are ambassadors of the regions in the central legislative process (there are other reasons as well, especially if a mixed electoral system is introduced; and in the way concurrent fields of competence are expressed), and agree with you and Sumane that representation in the second chamber cannot be on basis of equality of regions as in the US model.

I think that the discussion about the Second Chamber is absolutely crucial in regards to the ‘shared rule’ elements of a possible future Sri Lankan constitution. I would like to (1) make a short theoretical comment, (2) discuss some aspects of the Belgian system, (3) comment on the possibilities of shared rule in an asymmetric system and conclude with (4) the question of the powers of a Second Chamber.

1. If we go back to Lijpharts definition of consociationalism, he elaborates four different elements: (1) a government ruled by a grand coalition; (2) a mutual veto; (3) proportionality in recruitment to decision making bodies and public services; (4) and a high degree of autonomy for cultural communities. So if we are talking about consociational mechanisms having certain effects, it might be necessary to exemplify which of these mechanisms we refer to. Additionally, despite the fourth element of Lijpharts concept, all the others focus on shared and not on self rule.

2. The Belgian system is a unique polity in a young federal system. I do agree that since the last constitutional change 14 years ago, there have been a lot of tendencies that emphasized self rule rather than shared rule. Almost the only shared rule institution besides the federal government is the ‘concertation committee’. However, the reason for this is not only the communal arrangement of the federal structure (and to talk of secession is far reached). Rather, the territorial structure of the federation allows only for one identification scheme: besides the three cultural communities, there are only three territorial units. It might be worth to have districts that can serve as the basis to develop a further identity besides just belonging to one of the language communities. And thus, enhance the possibilities of cross-cutting cleavages.

3. In a situation of ethnic conflict, one can a.o. imagine either autonomy regulations or a fully fledged federal system. Anyhow, what possibilities of shared rule offers an asymmetric system? I can not think of any other than elitist institutions. If this is so, is it really possible to foster a common identity building? Additionally, when looking at the economic data of the country, a decentralization scheme with more powers given to the provinces in the South might be a good possibility also in terms of effectivity and efficiency.

4. I agree that the Second Chamber is an important institution for shared rule elements. Thus, I am convinced that there has to be some sort of overrepresentation (which would be marginal, as we see in the table). After talking about the composition of the senate, the more important thing is to talk about its powers. Otherwise every discussion about the composition is pointless. Here, the element of veto-powers comes into play, like the “alarm bell” mechanism in Belgium, where linguistic minorities can stop a legislative process, when they have the signatures of three fourth of their PMs.

Re your first point, I think it is important to underscore the importance of the difference between ‘consociationalism’ and ‘integrationism’ (Horowitz prefers ‘centripetalism’). Integrationist arrangements are concerned with building commonality whilst respecting diversity and rejecting pure majoritarianism.

Thanks for your response. Re the second point – I see the logic of your thinking (a) on the demographic basis of representation in the central second chamber; and, critically, (b) that you posit this on the assumption of some form of asymmetrical regional autonomy in the Northeast. This I think addresses my second critique well at the conceptual level. Naturally, I am fully with you on the need for entrenchment of the counter-majoritarian principle; indeed, in my view, the fundamental flaw in the ISGA is precisely this.

Thanks for your comments. Both your points are valid. I do not call myself consociationalist in a strong sense; my position is also centripetalist in the sense Horowitz uses the term. I totally agree with the importance of shared rule. In a way, my suggestion would promote shared rule in the sense that I have tried to underplay the territorial conception of power-sharing by privileging the cultural element, which in my view is less controversial. This is the idea that the Austrian Social Democrats advance in response to Lenin’s territorial assumptions about the self-determination of peoples. The Senate I proposed would be a constitutional forum in which the similarities of many cultures as well as the importance of difference (not differences themselves as an end in itself) would be respected.

I disagree with your second point, because the basis of my entire argument is opposition to majoritarianism at any level of government. In the scheme I propose, power-sharing would be based on the following demographics and logic (the figures are based on the current census):

On this basis, I believe the conflict-related issues regarding the North and East can be addressed through asymmetrical constitutional arrangements etc.

I am in full agreement with the desire to ensure respect for diversity and to have such diversity reflected in the Senate (as well as the institutions to be established by the constitution). I also find very interesting the attempt to spatially blend the communitarian and territorial bases of representation in the Senate, some form of which would be necessary in future arrangements.

However, (without the advantage of deserved reflection), I would make two preliminary observations that I think are potentially problematic for the schema proposed.

First, I think there is too much emphasis on constitutional mechanisms. I do think that some consociational devices are indispensable to our future constitutional dispensation, but would urge that they have to be balanced against other considerations, not least of which is efficiency of decision-making. To use the example employed, Belgium is a classic case of the constitutionalisation of difference to the exclusion of unity; the result of which is the post-constitutional framing of the dynamics of politics that focus wholly on self-rule to the exclusion of shared-rule. This has in turn reinforced the inefficiency of consociational decision-making procedures at the centre. The sum-total of all this is that Belgium is no longer perceived to be a viable State among its diverse citizens, who cannot see the point of the Belgian State. Of course in prosperous Western Europe, under the unique umbrella of the Eiropean Union, Wallonia and Flanders may separate without any of the violent consequences that will ensue in that eventuality in countries such as ours.

Other, more apposite examples of the perils of consociationalism are Lebanon (where arguably, constitutionally, the country is ungovernable even if the other geopolitical dynamics were absent) and the future Bosnia-Herzegovina (once the sunset clause on the Office of High Representative – the sole effective institutional link between the autonomous units – comes into operation).

Thus my argument is not about a wholesale rejection of consociational mechanisms, but that these should be balanced against other constitutional postulates such as effectivity and efficiency as well as the need to build an inclusive common identity. Put another way, consociationalism should be made use of for certain representation purposes, but it cannot be the fundamental organising principle of the future constitution. Therefore, my concept of consociationalism is narrower than what seems to be proposed.

Second, while I fully endorse the desire to ensure weighted representation to the non-Sinhala communities (including the laudable attention paid to the smallest communities), I would strongly argue that the particular schema proposed loses sight of the fundamental dynamic of the present conflict, as primalily between two nationalisms, one majoritarian, the other, minoritarian. In the schema proposed, Tamils of the North and East would be entilted to the same number of seats (5) as the Upcountry Tamils and the Muslims. I would argue that any post-conflict settlement to be viable would require more substantial and asymmetrical recognition of the Northeastern Tamil identity. This is not merely a realpolitik consideration of persuasion (i.e. ensuring ‘buy-in’)of a secessionist community into a shared constitutional compact. It is fundamental to constructing a Sri Lankan identity through power-sharing on the acknowledgement of the distinctiveness of the Northeast.