Differentiation theory and the ontologies of regionalism in Latin America

In this article, we argue that conventional understandings of regional integration based on neo-functionalism, hitherto often used to describe the diverse projects of Latin American regionalism, are of limited utility in that context. Rather than representing processes of economic or political unification, the various regionalisms could be understood more productively as a reaction to the crisis in legitimacy that social orders in the region have experienced under the conditions of globalized modernity. We then deploy an understanding of regionalism derived from sociological differentiation theory in order to advance this argument.


Introduction
T he processes collectively identified as Latin American regionalism are, in the academic discipline of International Relations as well as in political and media discourse, generally described using the terminology of regional integration based implicitly or explicitly on neo-functionalism.This application of a theory originally devised to explain the establishment and expansion of supranational political structures in a European context to a geographic area with very different conditions has been problematized before in the literature (Malamud 2010).Yet, what is lacking so far is a convincing theoretical alternative able to describe Latin American regionalism on its own terms.In order to arrive at such a description, we first retrace the debates surrounding contemporary regionalist projects.We come to the conclusion that the actually interesting puzzle is not so much the question of why those projects are behind the European Union in terms of supranationalism and institutionalization, as this can be easily Rev. Bras. Polít. Int., 60(1): e017, 2017 Kleinschmidt and Gallego Pérez 2 explained by conventional neo-functionalist theory.A more productive research question would be: If Latin American regionalism has made such little impact in political and economic terms, why are the diverse organizations highly politicized, with a strong presence in regional political debates?We then utilize an understanding of region and regionalism derived from sociological systems theory and its concept of a functionally differentiated world society (Albert and Stetter 2015) in order to undertake a re-description of regionalism and of the problems that it may contribute to solving in a Latin American context.The conclusion is that in the cases examined, regionalism is not so much based on the structural coupling of functionally differentiated systems, but rather on the generation of semantics to legitimize different state strategies for coping with the complexity of a globalized world society.We then discuss implications both for the future of Latin American regionalism and for possible comparisons with other regionalization processes in the Global South.

The puzzle of latin american regionalism
Latin American regionalisms and the organizational efforts derived from them are often a source of irritation for both academics and policy experts working in the field.The precise reasons for their criticism -or lack of comprehension, depending on the individual point of view -appear to be diffuse; they vary from case to case and from author to author.What most authors do share is the point of view that existing organizations have a lower degree of institutionalization than one could (or should) expect according to various criteria, and therefore depend mostly on the personalities of Latin American presidents and their personal preferences.Especially the more geographically inclusive and functionally comprehensive organizations, CELAC and UNASUR, have been criticized as consisting of, "little more than a series of roving summits for presidents and foreign ministers" that besides "empty symbolism" were capable of little else than collectively emphasizing national sovereignty at all costs, even over universal human rights norms (Sabatini 2014).Other authors describe the latter as a mere forum for discussions between national governments, which might provide a ready-made format for crisis management, but otherwise was not charged with any of the ambitions typically associated with regional integration projects (Briceño-Ruiz 2010, 227).
ALBA-TCP, the grouping of countries inspired by the late Venezuelan president Chávez, though celebrated by some as a "third-generation", outwardly oriented, progressive and inclusive "counter-globalisation project" (Muhr 2011, 99), is also evaluated negatively in terms of having achieved any of its stated intentions in relation to membership or policy.While the possibility of receiving Venezuelan cash and petroleum with no legal or fiscal controls might have attracted political elites of several countries in the region to what could be called a form of transnational clientelism (Kennemore and Weeks 2011, 272), it has essentially remained a pact of vaguely left-wing governments that seems to become increasingly unattractive as Venezuela's economy falters and its alternative development model appears to have brought few lasting benefits to its own citizens Rev. Bras. Polít. Int., 60(1): e017, 2017 Kleinschmidt and Gallego Pérez 3 or to those of other countries (Ortega and Rodríguez 2008).Proponents of counter-hegemonic integration criticize that ALBA has not evolved beyond an initiative of the Venezuelan government with few multilateral elements whatsoever (Lo Brutto and Vázquez Salazar 2015, 69).And while regional monetary arrangements such as the ALBA-affiliated international accounting unit SUCRE could in theory have the potential to provide a measure of independence from North American central bank policies and currency fluctuations, in practice, its economic relevance has remained insignificant in the context of markets heavily dependent on trade with the North (Trucco 2012).
The newest of the multitude of Latin American regionalist projects is the Pacific Alliance, which in ideological terms often is described as a counter-effort to ALBA and other more statist efforts at integration.Indeed, the emphasis of the PA seems to lie in trade liberalization measures and other policies highly compatible with liberal or 'Western' preferences in international politics.This has resulted in predictable enthusiasm from free trade adherents, who emphasize the potential for Latin America to act as a "rule maker" instead of a "rule taker" in international trade negotiations through the considerable economic weight of the PA (Rashish 2014, 3).While there might be enough reason to believe that coordinated action by Latin American governments could result in an overall stronger position in trade and other areas, this does not necessarily mean that the PA actually represents an outward-looking, institutionalized form of regional integration, understood as the pooling of decision-making capacities in a form of rule-bound organization possessing collective agency to some significant degree, nowadays almost exclusively associated with the EU (Söderbaum 2016, 171).Actually, current research on the political logic of the PA rather tends to demonstrate its strong connection to intra-regional "soft balancing" behaviour -either against ALBA or Brazil as a potential regional hegemon -as well as its reliance on national governments currently in office, and on their ideological preferences (Flemes and Castro 2016, 89).Like ALBA, it remains essentially a program contingent on the personalities in charge of member state executives and of their ideological affiliations.
Mercosur, of all the different projects examined here, in some aspects conforms most closely to the conventional vision of regionalism as political integration.It is often seen as the most ambitious, most institutionalized and might well be the most studied project of Latin American regionalism -or at least, it is the one project generally named as the definitive Latin American attempt at integration when comparing different regionalisms (Buzan and Lawson 2015, 302).And unlike most other efforts, it has actually had a significant effect in terms of trade diversion.Undoubtedly, many of the norms and structures that make up Mercosur closely resemble those of the European Union, generally following its "common market model" and even in some cases "literally copying EU terminology" (Lenz 2013, 215), or basing its parliamentarization initiative on the necessity of mirroring the parliamentary aspects of the EU (Söderbaum 2016, 189).These isomorphisms may have increased Mercosur's visibility and contributed to the relative importance accorded to it by external actors, including the EU itself.However, the parliamentarization process has not advanced further than other regional democratization endeavours such as the Andean Parliament or the Latin American Parliament.All three regional parliamentary bodies have purely Rev. Bras. Polít. Int., 60(1): e017, 2017 Kleinschmidt and Gallego Pérez 4 consultative functions while also serving on occasion as vehicles for "parliamentary diplomacy", while remaining irrelevant in terms of political decision processes (Mariano et al. 2017, 14).
Mercosur has therefore demonstrated the limits of "institutional mimesis", as the process has been aptly called, stopping far short of the political and legal supranationalism that characterizes the EU and its parliamentary component, and does apparently not seek to replicate it despite repeated claims to the contrary (Dri 2010).Its institutional deepening is now often considered to have stalled.Furthermore, the largest Mercosur economies, Argentina and Brazil, have over time developed quite divergent interests regarding the common external trade policy (Kono 2007, 179), and with new administrations that are not necessarily ideologically aligned with regionalist projects, might increasingly lose interest in them.In political terms, the strong limitations of Mercosur as an integration project became visible when member states used the controversial suspension of Paraguay for violations of the democracy clause in 2012 to -just as controversially -admit Venezuela, despite the latter probably being unable to comply with that clause as well as with other regulations.As a result, the legitimacy of Mercosur became highly disputed as a seemingly politicized instrument of arbitrary power politics (Marsteintredet et al. 2013), and currently, its future appears as being -just as in the other cases -highly dependent upon the political currents of the day.The fierce debate currently surrounding the expulsion of Venezuela due to violations of the democracy clause and non-implementation of treaties does, however, demonstrate Mercosur's relevance in regional public opinion as well as its potential impact on national political controversies (Gill 2016).It must be added, again, that these events are not an expression of some process of institutional evolution, but of the changed political orientation of the governments of Brazil and Argentina and the internal dynamics of Venezuela.

The eurocentricontology of latin american regionalism
Overall, the current literature highlights in virtual unison the very limited degree of regional integration achieved in Latin American politics.Important factors used to explain this perceived failure are, firstly, the strong degree of presidential centralization of power in the region -"the uncertainty of interstate cooperation is linked to the uncertainty of the executive office itself, continually susceptible to short-term visions and the electoral cycle" (Emerson 2015, 497) -and secondly, the almost completely missing security dimension of regionalism in a continent without significant interstate conflict or even offensive military capabilities (Dabène 2009, 7).Finally, and perhaps most significantly for neo-functionalist approaches that mostly understand regional integration as economic integration, the lack of important trade relations between most Latin American countries is cited as a major factor: whereas most highly institutionalized integration schemes were built on a structure of pre-existing regional economic interdependencies, Latin America could be characterized rather by its extra-regional participation in global trade circuits, typically in the role of commodity supplier.Any trade-related benefits that might be achieved by actual integration would be small and anyway reaped mostly by Brazil, the "highly cost-averse leadership" which would reduce any remaining interest on behalf of the other states (Burges 2005, 451).For these and other mostly secondary reasons, states in the region therefore had a tendency to "speak regional, act unilateral, and go global" (Malamud and Gardini 2012, 131).
One could probably add to the list of unmet expectations by including other regionalist projects, such as explicitly sub-regional approaches like the Andean Community of Nations (CAN) or the Central American Integration System (SICA), or by also examining the trajectories of past attempts at integration that often have been subsumed under or superseded by more current projects, such as the Latin American Free Trade Area (LAFTA).We shall concentrate here on those cases that are both contemporary and represent -at least in terms of their general approach -a model for an entire region, whether that region is considered South or Latin American.A more complete overview is given, inter alia, by Olivier Dabène (2009).In all those cases, the long-standing expectation that a region with a significant common cultural and political heritage would find it comparatively easy to establish substantial forms of integration -in fact, an expectation going back at least to the Carta de Jamaica written by Simón Bolívar during the South American wars of independence, affirmatively referenced in a majority of the region's constitutions (Herrera 2016), and at least cautiously shared by theorists of regionalism during earlier waves of Latin American attempts at integration (Haas and Schmitter 1964;Seligson 1999, 129;Muhr 2011) -has been disappointed regularly, only to re-emerge just as regularly: "Latin America in particular has a long history of truncated regional aspirations" (Riggirozzi 2012a, 428).
These pessimistic diagnoses should not be taken to mean that Latin American regionalisms have had no effect at all -after all, especially Mercosur has established stable patterns of cooperation between erstwhile rivals, which have touched critical spheres such as nuclear energy (Darnton 2014).In some issue areas, limited progress in the direction of integration might have been made -examples to be mentioned include the Court of Justice of the Andean Community, which has become a strong influence in sub-regional legal systems (Howorth 2012, 146), the significant role of UNASUR as a vehicle for the promotion of "a right-based approach to health" (Riggirozzi 2012a, 419) or the strengthening of mutual understanding in the negotiation processes of agreements (Prieto 2016, 304).Of course, these might be seen as small-scale victories in a much larger struggle for 'real' integration.In general, however, it is notable that those aspects of Latin American regionalist projects that have been described as remarkably successful by scholars are otherwise mostly ignored in academic discussions, which tend to focus on the question of supranational integration, and are also not a cause for public debate.
Yet, the most plausible step towards integration in the conventional sense would lie in a targeted expansion of those regional activities that actually have managed to generate consensus and some degree of measurable success, and try to base new institutions on their established, functioning mechanisms.Such processes were at the core of the development of regional integration in Europe, and are observable in a number of regions today (Howorth 2012, 148).Indeed, some Latin American integration projects seem to be making advances along those same lines.Yet, such Rev. Bras. Polít. Int., 60(1): e017, 2017 Kleinschmidt and Gallego Pérez 6 progress is not what the conversation is focussed on; rather, it is dominated by the insistence on the possibility of rapid and complete integration to such a degree "that anything short of integration is a political failure or, worse, a betrayal" (Malamud and Gardini 2012, 121).This might be compatible with the assumption that in the regions of the Global South, the desired outcome of regionalist efforts is actually the political and economic "autonomy" of states, and not so much integration per se (Acharya 2016, 117).This would explain a number of features of regional organizations in Latin America.However, this assumption also generates the question why precisely those components of the latter that would contribute most to national autonomy are those that lag most, and have, such as in the case of the "Banco del Sur", not been implemented despite agreement on the basic framework (Grugel et al. 2008).
In general, the existing literature posits the failure to institutionalize some form of supranational decision-making in regional organizations as the explanandum.The latter, although "meant to be an expression of unity and solidarity", have turned out -or rather, have quite regularly turned out -to be "a stark reflection of diversity and heterogeneity" (Gardini 2012, 88).This is regularly posited as the puzzle of regionalism in Latin America, which also is described as being, besides Europe, "the 'other' continent with a long tradition of modern regional integration" (Dabène 2009, 3) and integration having "been sought by Latin American countries for decades" (Briceño-Ruiz 2014, 1).We find this emphasis interesting -because, as described above, the continuing failure to establish classical, EU-style supranational integration is actually quite well explained within current models of integration theory (Mattli 1999, 146).Rather, the questions that we actually do find puzzling are why, after decades proclaiming lofty goals of integration, do Latin American countries still spend significant political and symbolic capital on efforts that are ultimately judged as ineffective by the standards of supranational integration theory?(Marsteintredet et al. 2013).Why do embryonic efforts at regional integration enjoy massive public attention, to the point of enjoying a somewhat similar politicized status in public opinion as the comparatively, in material terms, much more materially consequential topics of EU decision-making?(Hurrelmann and Schneider 2015, 253).Is it not likely that the observable regionalist projects, even though they sometimes appear to mimic European efforts, operate within an entirely different logic?And how could that logic be elucidated?
We believe that research on Latin American regionalism ultimately must go beyond the focus on economic integration, supranationalism, autonomy, and other approaches based upon the "state-as-container" model (Taylor 1995, 1), and that the continued insistence on this focus in academic discourse very often is due to the influence of neo-functionalism as an almost hegemonic intellectual model in the definition of regional integration.Yet, neo-functionalism has a significant weakness in being virtually unable to address any regionalist model other than supranationalism, which it posits as the generally and universally most efficient form of integration (Mattli 1999, 10).Yet, many of the causal factors described by neo-functionalist authors cannot be considered as particularly relevant in our context.The integration of "key strategic economic sectors", as in the European Coal and Steel Community (Rosamond 2000, 51), will be of little relevance when most important export industries are extractive, linked to markets in industrialized countries.Where potential might exist, such as in the creation of a "peripheral automotive space" for the Brazilian and Argentinean automobile industries via Mercosur, the benefits have been distributed unequally -resulting essentially in a protected industrial sector for Brazil, with a more peripheral role for the other member states (Obaya 2014).Neo-functionalism's transnational technocratic elites administrating regional institutions and pushing for more integration, where they exist, have in most cases not acquired an independent role in the formulation of policy in highly executive-centred political systems.More importantly, if they were working under ideal conditions, they would probably emphasize extra-regional engagement, as this is where most relevant economic connections are made (Malamud and Gardini 2012, 129).The economy-driven spillover automatism assumed in neo-functionalism simply has no empirical basis.
Therefore, neo-functionalism actually provides a sufficient explanation for the virtual absence of supranational forms of regional integration in Latin America.What it cannot do is explain any other dynamic of regionalism, which appears necessary in the case of Latin America -and perhaps of other regionalisms as well.This does not imply that conventional theories of regional integration have no relevance at all for Latin America, or that they could not become more relevant if certain conditions change in the future.Specifically in the judicial realm, some regional court judges appear to display a higher degree of autonomy vis-à-vis member states than other institutions (Alter and Helfer 2010, 586).But overall, it is far from clear that future developments will play out according to the neo-functionalist script.Assuming that it will, and evaluating Latin American regionalism accordingly, it might generate "a false image of homogeneity because it presupposes equivalent phenomena" (Mariano et al. 2017, 3).While there can be little doubt that the academic and political vocabulary of contemporary regionalism has been shaped heavily by concepts developed for the understanding of the EU, it has recently been argued forcefully that studies of regionalism in IR must seek concepts not based on the implicit example of the former (Acharya 2016, 109).At the same time, other scholars have warned to not fall into the trap of elevating claims of subaltern knowledge and "Southern solidarity" to such heights that they obscure the relevance of global structures and dynamics (Hurrell 2016, 10).In this paper, we try to satisfy both desiderata by applying the world society theory of Niklas Luhmann to our puzzle, which recently has been established as a novel possibility for innovative theory-building in IR (Albert et al. 2013;2010;Albert 2016), and has already been applied to the problematique of concepts of regionhood (Stetter, 2008).

Functional differentiation as the context of regionalism
Unlike traditional ontologies of regions, such as those underlying classical IR theories of regionalism and integration, our preferred approach, the sociological systems theory established in the works of Niklas Luhmann (2012)  by Mathias Albert and Stephan Stetter (2015), is not based on primordial classifications of territorial spaces or culturally or racially coded groups of human beings.Rather, social systems are defined by the processing of meaning within communication.Humans as biological or psychological entities are not considered component parts of society -even though they represent its relevant and necessary environment, which is why the self-description of society as a geographically delimited aggregation of human beings sometimes may function as a useful descriptive device (Luhmann 1995, 210-4).This would be the case as well in traditional conceptualizations of what a region generally is considered to be in regionalist IR approaches.Another difference in relation to classical container models is that, according to Luhmann's theory, contemporary society must be understood as a world society, containing any and all communications that can potentially connect to each other.Society is not a priori separated into territorial blocs such as nations or regions, as traditional European concepts of society maintained (Luhmann 2012, 83-4).Contemporary concepts of regionalism tend to continue these traditions.In their place, world society is described by its primary form of differentiation -the latter being defined as the principle conditioning mutual connectivity of communications.Interactions at the micro-scale can establish rules of limited temporal and topical range based on ad hoc consensus, or organizations might synchronize communication at significantly higher degrees of complexity, but modern society as a whole must be described as based on a much more general level of rules for potential connectivity (Luhmann 1995, 154-5).
The essential principle establishing those rules in modern world society is considered to be its primordial differentiation into operationally closed function systems.These produce and re-produce themselves by generating boundaries of meaning based on system-specific binary codes -the distinctions of powerful or powerless in politics, payment or non-payment in the economy, legal or illegal in law (Luhmann 2013a, 53), to name just a few -which allow communications to connect to other, similarly coded communications in a self-referential, iterative process (Luhmann 2012, 297).Modern world society thus distinguishes itself from earlier historical structures by separating its internal elements in terms of referent problems for generating rules defining which communications can meaningfully connect to other communications, not in terms of stratification (as in class-based societies), of centre and periphery (as in ancient city-states or world empires), or of undifferentiated, segmentary units (as in tribal societies) (Luhmann 1998, 4).These historically older forms of differentiation may remain as secondary structures, but are always generated in relation to the distinctions established in function systems.In the modern world, we accordingly encounter system-specific hegemonic concepts such as that of the sovereign state as a semantic for the operational closure of the political system, of the market for the economic system, of academic freedom for the system of science, of positive law for the legal system, of abstraction or at least of self-referential art for the system of art (Luhmann 2000, 149), and so on.
Diverging from more traditional assumptions about societal structures found in IR theory, differentiation theory does not accord any sort of ontological status to countries -these might constitute segmentary differentiation within the global political system -nor to inequalities in terms of different positions on global capitalist chains of production (as in world-systems or dependency Rev. Bras.Polít.Int., 60(1): e017, 2017 Kleinschmidt and Gallego Pérez 9 theory) -which would constitute a semantic describing centre-periphery differentiation within the global economic system.Also, any primary ontology of world regions is notably absent -"it is no longer plausible to think of Brazil as constituting a completely different society from Thailand" (Lee and Brosziewski 2009, 54), meaning that whether Brazil or Thailand are part of the same social system depends not on their geographical location, but on their significance established within communication.Regions and regional organizations do not exist per se, but must be considered to change their relevance based on functional context.For example, Mercosur has been observed through economic 'lenses' by most member countries, but as part of a more political project by Brazil (Teixeira 2012, 124), probably contributing to the conflicting interpretations described above.
Most general accounts of regionalism explicitly or implicitly assign priority to distinctions specific to one function system within a scheme of first-order observation -typical candidates are politics, the economy, or science, as in some post-colonial perspectives (Hurrell 2016).They thus preserve a conceptual apparatus from classical sociology based on ontological nationalism -or, rather, in our case, ontological regionalism: one problem -be it economic dependence, political disunity or intellectual coloniality -can stand for the whole, can be directly observed, and might be solved or ameliorated through political processes such as initiatives towards regional integration.Differentiation theory, in contrast, relies on second-order observation (Luhmann 1995, 300).It describes modern world society as dominated by problems that derive from the "conflicting rationalities" of various function systems (Kessler 2012, 78).Any occurrence within global society as a whole needs to be observed within the context of the multiplicity of systemspecific distinctions.No single functional system can represent a singular 'society' or 'region' as a whole, as they could in the primordial geographies underlying the theories that still dominate the field of study.This would also be a reason why models of regionalism that depend on a previous consensus on essential ethical or procedural norms, such as those of Held or Habermas, regularly encounter "a tension between the particular contexts of democratic legitimacy and the universalism demanded of a transnational or even global political culture" (Lupel 2005, 119).
The traditional ontologies of Latin American regionalism are not based on the assumption of a functionally differentiated world society, but on semantic traditions of geographically delimited societies.They assume that economic or political motivations are at least potentially capable of providing conceptual order to overall societal structures.Perhaps the most widespread expression of this notion in Latin American IR is found in the concept of "autonomy", which often stands for the postulate that container-states as unitary actors can effectively manage the diverse functional aspects of society within their geographic containers in order to make them more independent from the outside world (Olaya Barbosa 2007).The conventional ontologies of regional integration considered above generally presuppose an equivalent chain of causality, only transferred to a higher spatial scale.This is not extraordinary by itself, as numerous social contexts as well as established academic disciplines rely on such a "hypostasis" (Bommes and Thränhardt 2010, 32), observing any and all social facts exclusively through the prism of one function system.In some cases, such styles of observation provide a perhaps useful reduction of complexity.
Rev. Bras.Polít.Int., 60(1): e017, 2017 Kleinschmidt and Gallego Pérez 10 This is possible temporarily, as functional differentiation as the deep structure of society on one hand and its generalized description in the form of societal semantics on the other may diverge significantly until that divergence creates implausibilities that constructivist IR theorists might refer to as a condition of "ontological insecurity" (Mitzen 2006, 345).In our theoretical context, this means: established semantics will need replacement when they become notably implausible for describing the form of differentiation to which they are related (Stäheli 1997, 131).To give a well-known example, Waltzian structural realism made explicit mention of functional differentiation in its Durkheimian version, but restricted it to domestic society, performing the above-mentioned hypostasis in favour of a primacy of politics.In IR theory, only in the post-Cold War era and the ensuing debates on globalization and transnationalism could the primacy of functional differentiation be plausibilized (Zürn et al. 2013, 235).Nowadays, even IR theorists outside world society theory allow for a significant role of functional differentiation (Donnelly 2012), and some go so far as arguing that any approach that does not recognize the dominance of functional differentiation in modern society could be summarily dismissed (Buzan 2014).For theoretical accounts of Latin American regionalism, this moment has probably arrived.As demonstrated above, conventional narratives do not seem to go much beyond the diagnosis that regionalist efforts are either simply irrelevant, or at least do not match the pattern established by European integration, but without providing an alternative account of what their relevance might be.

Peripheral modernity and latin american regionalism
Applying systems-theoretical terminology to the neo-functionalist work on regional integration, the latter also hypostasizes the primacy of a function -in this case that of the economic system, which is considered to be an almost automatic driver of supranationalization processes (Rosamond 2000, 52).The causality is strictly one-way, the logic of one system is a means to the ends of another system's logic.A systems-theoretical approach to the study of regionalism, based on the primacy of functional differentiation, would rather emphasize, in the case of the EU, its role as "a coordinator and arbitrator between functionally differentiated systems" (Kjaer 2007, 374-5).More specifically, the EU fulfils the needs of various function systems to reduce the complexity of their legal, economic, and political environments by homogenizing, mostly within and through the system of law -a necessity which "becomes more pervasive the larger the overlaps between similar dynamics across various function systems are" (Albert and Stetter 2015, 79).Most importantly, it "limits the autonomy of the political and legal systems of the member states" in a way that safeguards the operational closure of several function systems -of law and politics by reframing them on a larger geographical scale, ensuring the relevance of their distinctions vis-à-vis other systems, and of the economy and other systems by, at the same time, narrowing the parameters within which politics and law can operate.Regional integration, understood in this way, is not the opposite of globalization or of world society.Rather, it functions as their essential enabler Rev. Bras. Polít. Int., 60(1): e017, 2017 Kleinschmidt and Gallego Pérez 11 under the conditions of extremely dense couplings between function systems (Kjaer 2007, 378), which IR theory would -in the case of politics and law -classically describes as interdependence.Does this imply that contemporary Latin American regionalisms are nothing but 'EU-in-waiting', waiting, perhaps, for some future growth in world societal complexity to become necessary and active?Is the systems-theoretical interpretation of integration, then, nothing but a neo-functionalist account enriched by social theory, and the linearity assumed by neo-functionalism still the only way to compare the two?Such a point of view would conform with the "tendency […] to segregate and stratify people and things on the basis of their quotient of 'modernness'.'Modernity' operates as a qualitative, as much as a chronological, category" (Sharman 2011, 491).Such diagnoses inform the identification of Latin America with the Global South or other geographies of subalternity, which often serve as a conceptual background for descriptions of the corresponding regionalisms.But if modernity refers to the worldwide primacy of functional differentiation, this cannot imply that "peripheral modernity" (Neves 2001) simply follows a different logic of differentiation.More or less modernity, under the primacy of functional differentiation, cannot be described in terms of stratification or of centre and periphery as historic forms of society -even though the notion of Latin America as a "living museum" has provided a powerful metaphor for such assumptions (Ebel 2003, 25).On a superficial level, the reference to the persistence of highly unequal, status-based, or patrimonial structures might appear plausible.Yet, those phenomena cannot, in our theoretical framework, simply lead to a classification of Latin America as somehow 'not modern'.There is no "common difference schema" that would automatically make such a classification viable for an overall description of inequality in world society theory (Luhmann 2013b, 88), as more traditional terminologies might imply.The role of Latin America cannot be described as belonging to a generalized, unspecific periphery that remains somehow isolated from functional differentiation.After all, many of the aforementioned, seemingly atavistic phenomena are actually produced and preserved by distinctions established within global function systems, such as the profitability of resource extraction or of illicit narcotics trafficking, and they not only coexist, but are very often co-constitutive with the 'normal' functioning of those systems (Maher 2015).
Rather than tolerating historic curiosities, world society in its peripheral areas develops different mechanisms for addressing -or 'including' -individuals in the context of function systems (Neves 2001).The 'normal' functioning of modern world society assumes that individuals can be addressed by communication in function-specific terms -for example, creditworthiness in the economic system, due process in the legal system, qualifications in the education system, the power to vote in the political system, and so on.In peripheral modernity, however, those function-specific modes of inclusion and exclusion are displaced by a generalized form of the latter distinction (Luhmann 2012, 98).Again, this does not represent a displacement of functional differentiation, but rather often a consequence of the latter: The lack of a (legal) title to land leads to the non-payment of taxes, which leads to a lack in local public infrastructure, which leads to the creation of informal power networks -or, conversely -the availability of money through resource extraction leads to the possibility of vote-buying, which leads to formal political influence, which leads to the legalization said resource extraction, and so on.These and other imaginable forms of "chain exclusions" (and corresponding chain inclusions) appear to be a general characteristic of social structure in the Global South (Stetter 2008, 105-115).
Therefore, unlike in 'centric' modernity, peripheral forms of inclusion and exclusion lead to a "miscellany of codes and criteria of communication" being utilized all at once to process not functionally specific distinctions, but a generalized distinction of inclusion/exclusion (Neves 2001, 243).And this is where regions might come into play in world society theory.In the currently emerging differentiation-theoretical approach to regionalism, regions "are seen as social forms resulting from a specific combination or interaction of forms of differentiation" (Albert and Stetter 2015, 72).In other words, regions in a functionally differentiated world society must be described in terms of their relation to functional differentiation as their primary characteristic.In terms of social semantics, in peripheral modernity, regionalism therefore tends to take the form of explaining group-based, function-unspecific, inclusions and exclusions of groups of humans, which may be highly idiosyncratic, considering for example the sectarian distinctions prevalent in parts of the Middle East (Stetter 2008, 122).In more general terms, the relation of regional semantics and world society may be described as such: The distinction and mutual interplay between universalistic and particularistic approaches to sociological knowledge and the legitimization of norms have their empirical correlate in the distinction between world society and region: particular regional features cannot be delineated as if isolated outcomes of endogenous evolution.All regions may now need to consider themselves, under the contemporary conditions of factual, social, and temporal interrelationships, as a moment of modern world society.Even if we cannot pursue it here, it is our assumptions that similar processes of structure-building in different regions of world society lead to different yet equivalent semantic responses (Mascareño and Chernilo 2009, 75).
It is possible that Latin America, of all world regions, finds it more difficult to produce semantics that can explain and relate processes of chain exclusions and group inclusions to functional differentiation.The Middle East, as the only world region studied so far under these premises and in terms of differentiation theory, apparently relies on a body of religious, familial and colonial norms that can relate problems of inclusion and exclusion to function-specific distinctions (Stetter 2008, 134).One might speculate on possible roles of claims of state sovereignty and/or civilizational othering for the East Asian region and for the post-Soviet space.Latin America, possessing no semantic tradition of its own that could fulfil this function -having adopted Western notions of "homogeneous universal progress and civilization" already during its process of political independence (Mascareño and Chernilo 2009, 75) -appears to define its regional structure of inclusion and exclusion more in terms of isolated problems, often described in terms of the abovementioned atavisms, without an overarching semantic that enabled it to communicate them in adequate terms.
Yet, structures of generalized inclusion and exclusion that could constitute a specific 'regionness' in the sense outlined above, and that display the characteristic blending of functional distinctions, Rev. Bras. Polít. Int., 60(1): e017, 2017 Kleinschmidt and Gallego Pérez 13 can easily be described.The difference between political and economic communication, for example, may be void within structures that could be called clientelistic, subverting in many cases also the typically modern left-right difference of the political system (Ruth 2016).On a wider scale, but similar in functional terms, the provision of state services and infrastructure is dictated often not by political programs and decisions, but by private interests related to commodity exploitation (Saylor 2012).Academic research and debate is often understood not in terms of disputing explanatory or interpretative propositions, but as a contribution towards a political project or towards generalized positive societal change in order to achieve inclusion in politically guided processes of resource distribution (Mu and Pereyra-Rojas 2015).More tragically, law enforcement is often associated with the excessive use of violence -not in order to generate security or fulfil a legal imperative, but to provide some measure of authority for police agents in the absence of a class-based differentiation from the public in low-income neighbourhoods (Garmany 2014).While a more extensive study of inclusion and exclusion in Latin America must yet be undertaken in systems-theoretical terms (for the case of Mexico, see Millán 2008), the examples listed here suggest as a general theme that the distinctions utilized are based mostly on distinctions of economic class closely tied to, but not identical with, the political power of the state.It is hardly surprising that it were these elements that were used post-1945, in the context of dependency and world-systems theory, to construct some sense of regional identity for Latin America.
Classical accounts of world politics often tend towards the assumption that world order should be described in terms of an overarching conflict that encompassed political as well as economic, legal, and other aspects (Stetter 2014).Latin American semantics of regionalism during the Cold War closely followed this script of identifying social order with aprimordial structure of conflict that could fuse function-specific logics into two opposing sides (Luhmann 1995, 390).Existing conflicts, especially in South America, gradually were subsumed into a form of regional intergovernmental solidarity against possible communist subversion (Buzan and Waever 2003, 313).With the narrative of the Cold War coming to an end after 1989, conflicts all over the world became localized affairs that were unable to connect to a globally relevant semantic (Stetter 2014, 51-2).With the end of the Cold War, a global conflict that would supersede functional differentiation was no longer a plausible semantic -not even in the "global war on terrorism" that dominated headlines for some time after 2001, mostly due to its lack of a debatable political issue being at stake (Buzan 2006).For Latin America, its effects were negligible, except for some forms of security cooperation with the United States, which could claim new sources of legitimation, and new grounds for critique (Tickner 2015).What was lacking was a semantic to express structural deficiencies of the characteristic form modernity had taken, shared to varying degrees throughout the region, and to formulate normative claims based on them.
One study on regionalism in Latin America describes several on-going projects as being concerned less with political decision-making or economic benefit, but far more with regional "identity formation", that is, with the "redefinition of regional consensus over social and economic resource sharing" (Riggirozzi 2012b,425) gains a new meaning.The theory of functional differentiation allows us to re-describe this finding in a possibly productive manner: regionalist projects in Latin America have their function in the generation of semantics that enable specific structural features of inclusion and exclusion to be communicated in terms that are compatible with the codes of functionally differentiated systems in world society.In many cases, the ambitious programmes outlined in official documents serve as what has been termed "pre-adaptive" semantics -they occur "(prior) to, compensatory for, or co-constitutive of social-structural change" (Jaeger 2009, 130).This fits well with a study that examines social and political change in Latin America through the more traditional lens of either domestic or international causation, but ultimately describes the most salient problems as emerging from the different demands from within different "issue areas" (Arceneaux and Pion-Berlin 2005, 205-18) that systems theory would identify as pertaining to different function systems.The debates surrounding regionalist projects should thus be considered a process of "societal constitutionalism", allowing for the "political problematization of social inclusion/exclusion" (Holmes 2011, 132) -perhaps not too effectively and certainly polemically, but in a form that is highly compatible with worldwide criteria for political legitimacy, and gaining legitimacy from the involvement of states whose citizens share many or most related concerns.
Within this process, the Pacific Alliance seems to stick closely to the universal modernism of the era of independence wars, and identifies the problem of exclusion primarily with insufficient degrees of economic freedom.Mercosur might take an equally modernist approach, but one using the system of politics as its primary point of reference for the inclusion of individuals.Both, in their respective treaty texts, therefore make mutually opposed arguments over the primacy of the political and the economic system in the inclusion of individuals in modern society.Both do so in the form of "hortatory principles" in what may be considered an incipient debate on transnational constitutional principles (Kratochwil 2014, 112).ALBA, on the other hand, appears to communicate the chasm between expectations and reality in terms of the rejection of norms inherent in functional differentiation, relying instead on semantics of a morally based, not function-specific inclusion of individuals in a manner that remind one of the semantics of protest movements (Luhmann 2013b, 157).What they have in common -and what differentiates them from the EU -is that they do not actually have strong structural effects on the operation of function systems, or on their structural couplings.Instead, they produce competing sets of semantics that might be suggestive of alternative strategies for dealing with the complexity of a functionally differentiated world society.
Therefore, the regional divisions created by semantic divergences should not be portrayed as weaknesses, as conventional analyses are wont to do (Daniels 2015).Rather, they enable the debate and perhaps the gradual establishment of norms on how forms of social inclusion and exclusion can be processed in the regional context.If legitimacy for certain norms and courses of action thereby is generated, legitimacy as the "contingency formula" of the political system would enable a more effective making of political decisions (Luhmann 2012, 282), and might allow an important step in the future evolution of Latin American states, perhaps beyond the hard Rev. Bras. Polít. Int., 60(1): e017, 2017 Kleinschmidt and Gallego Pérez 15 dichotomy of inclusion and exclusion that often characterizes them so far.How exactly such a process might play out is, so far, indeterminate.More research on the negotiation processes of the norms generated in the various integration projects is needed to come to a more solid conclusion.Yet, our approach, based on differentiation theory and the distinction of structure and semantics, can elucidate why Latin American integration is, on one hand, relatively inconsequential in 'material' terms, but hotly contested in the political arena and considered as important by the states involved in them -exactly what was described as the puzzle of Latin American regionalism in the introduction.

Conclusion
By describing Latin American regionalism as a reaction to the semantic exigencies of functional differentiation as the defining principle of world society, we have been able to describe a generative principle of regionalism as pre-adaptive semantics that may provide a coherent interpretation of so far unexplained features of regional integration projects.Furthermore, it refers to a consistent structural logic that allows for transregional comparisons and is, at the same time, sufficiently distanced as to not be susceptible to accusations of eurocentrism (Albert 2016).Based on the arguments we made here, it appears as plausible that expectations of an eventual convergence of Latin American regionalism with the model of the European Union were based not on careful consideration of the functioning of Latin American politics and society, but on more impressionistic ideas, such as the much criticized vision of Latin America as a "second Europe" (Briceño Guerrero 2007, 37).Rather than constituting integration of political decision-making or resistance against Washington consensus-type policies, as it is described sometimes (Riggirozzi 2012b) -which for integration projects in Latin America would be doubtful in terms of intention in some, and in terms of effectivity in all cases -the function of regional integration in Latin America, according to our approach, lies in what constructivist IR scholars might call "ontological security" (Mitzen 2006), or what systems theorists would call an adequate reduction of environmental complexity.The future development of the different regionalist projects might therefore be very different from expectations generated in the context of neo-functionalist integration theory.Based on the successes and failures of different claims to legitimacy, one can probably expect a shifting, variable geometry of different competing projects rather than a continuous deepening of EU-style integration.But that might be not a bug, but rather a feature of a process that is fundamentally different in nature from European regionalism.
Of course, this article is mainly of a conceptual nature, and firmly establishing systems theory and functional differentiation as a viable perspective on Latin American regionalism will require more empirical engagement with the detailed workings of integration projects.Another necessary next step for research in this area would probably take the form of an exercise in comparative regionalism.After all, studies of other regionalisms in the Global South, such as in the Caribbean, and recently applied in theoretical IR research on regionalism . In the light of everything said above, this diagnosis