Between Autonomy and Dependency: the Place of Agency in Brazilian Foreign Policy∗

The article examines the construction of the concept of autonomy in Latin America and discusses to what extent it can be applied to contemporary Brazilian foreign policy. The article irst examines classical de initions of the concept, and then looks at the ways in which it has been used to analyze Brazilian foreign policy for over half a century. We then reaf irm the importance of agency and how power relations vary from one thematic area to another. In doing so, the article advocates the concept’s applicability for explaining certain behaviours, but rejects its use as a “grand strategy”.

Furthermore, while we concede that there are national particularities in the formulation of the concept (SARAIVA, 2014b), we do not share the belief that these national particularities can be part of the explanation for why, to take the example of Brazil, foreign policy exhibits an almost perennial aspiration for autonomy (SPEKTOR, 2014) by whatever strategy, without ever distorting the original meaning of the concept.
In methodological terms, our re lections are based on principles drawn from two strands of intellectual history (although these are not fully embedded in the ield): the linguistic contextualism of Quentin Skinner (1988), and Reinhart Koselleck's (1992) history of concepts.Without going so far as to propose "that it is only possible to understand the meaning of any given text, or even of an utterance or idea by 'recovering' the intentions of the author in the act of writing and by 'rebuilding' the context of the linguistic conventions available at a given historical time" (JASMIN, 2005, pp. 31-32), we argue of the need for greater analytical rigour in interpreting the guidelines of Brazilian foreign policy.Thus, just like Skinner (1988), we reject the thesis of timeless ideas; but, like Koselleck (1992), we also reject the assumption of their immutability.In any event, without going into the debate about whether or not it is possible to talk about a theory of autonomy 2 , it cannot be denied that in order to formulate the concept of autonomy, its founding fathers, Juan Carlos Puig (1980) and Helio Jaguaribe (1979), had to develop some very speci ic postulates.Therefore, in line with Koselleck's (1992) proposition, we argue that to apply the label of 'autonomist' to certain guidelines of Brazilian foreign policy without adherence to the postulates that originally constituted this concept, actually means the development of a new concept that is distinct from the original and which is articulated to other contents, even if the word used -autonomy -may be the same.In this sense, our objectives are similar to those of Ruiz and Simonoff (2017), who sought to "explore what is the explanatory power of these two authors' [Puig and Jaguaribe] proposals in a world that is surely different from the one at the time of their formulation" (RUIZ and SIMONOFF, 2017, p. 71).Moreover our main goal is to draw attention to the misconceptions incurred by recent attempts to use the concept of autonomy in describing Brazilian foreign policy guidelines, without adequate attention to the concept's key postulates.
That said, this article is organised in three parts, in addition to this introduction.Firstly, we examine classical de initions of the concept of autonomy in the Latin America context.Secondly, we analyse how this concept has been used to describe Brazilian foreign policy guidelines for over half a century, often lacking the rigour with which the construction of the concept itself bene ited.Thirdly, we reaf irm the centrality of agency in the classical formulation of the concept of autonomy, and how power relations vary from one thematic area to another.Rather than another attempt to rede ine the concept, we propose the thesis of the contextualization of power, adding the heuristic potential of agency bpsr Between Autonomy and Dependency: the Place of Agency in Brazilian Foreign Policy to explain social and political relations.This contributes to a more precise use of the classical concept of autonomy to explain certain behaviours, but stresses its inadequacy as generator of a "grand strategy", at least in the case of contemporary Brazil (RUSSELL and TOKATLIAN, 2015).Finally, we summarise the core points of this argument and underscore its main elements.

Origins of and challenges to the concept of "autonomy"
The irst and most robust de initions of the concept of autonomy in Latin America were elaborated by two renowned intellectuals, Helio Jaguaribe (1979) and Juan Carlos Puig (1980).Although they never collaborated, they have so many points in common that, according to Ruiz and Simonoff (2017), "it is possible to list them as part of a similar research project, in the Lakatian sense of the term" ( RUIZ and SIMONOFF, 2017, p. 82) .
We start from the same basis as these founding fathers, in order to discuss the concept's adequacy for analysing the recent trajectory of Brazilian foreign policy.In other words, this starting point provides us with the tools to engage with claims about the autonomist character of Brazilian foreign policy during certain periods of the country's history.
Both authors start with the assertion that the international order is hierarchical, and not anarchic as posited by realist theses (JAGUARIBE, 1979;PUIG, 1980).In this sense they identify the distinctiveness of those who speak from the periphery, and therefore experience the condition of dependency and the effects that characterise it.
For Jaguaribe et al. (1969), "autonomy (...) means, at the national and regional level, both the availability of conditions that allow free decision-making by individuals and agencies representative of the system, and the deliberate resolution to exercise those conditions" (JAGUARIBE et al., 1969, p. 66).Similarly, for Puig (1980), autonomy represents "the maximum capacity of choice that one can have, taking into account objective realworld constraints" (PUIG, 1980, p. 149).
There are components of structure and agency in both de initions, making the concept more robust and operational, as we shall discuss below.
With regard to structural components, these are divided into internal and external.In the irst case, it is about what Jaguaribe (1979) calls "national viability", i.e. the domestic conditions that allow -but do not guarantee -autonomist behaviour (JAGUARIBE, 1979).These are: the possession of adequate human and natural resources, the country's capacity for international integration, and its degree of sociocultural cohesion.In a similar vein, Puig (1980) considers the existence of suf icient material bene its to develop a national project and to put it into practice, and the explicit commitment of the elites to the same project as the key criteria (VIGEVANI and CEPALUNI, 2011, p. 31).Another key element Puig (1980) identi ies for the possibility of an autonomist project was the existence of a model of internal development and strategic solidarity with other countries that aspired bpsr Letcia Pinheiro & Maria Regina Soares de Lima to the same goal (PUIG, 1980, p. 155).Note that even among the structural determinants there is reference to choice, i.e. to agency.As we shall see, this is also the case for Jaguaribe (1982).
With respect to "international permissibility" -an external, structural component according to Jaguaribe (1982) -, this is characterised by the existence of "conditions to neutralise the risk from third countries endowed with suf icient capacity to exert effective forms of coercion" (JAGUARIBE, 1982, p. 22).Interestingly, this de inition, which for Jaguaribe (1982) is structural, involves not only economic and military capabilities, but also the adoption of a strategy for action, i.e. the "establishment of alliances" (JAGUARIBE, 1982, p. 22) -and so, in our view, is closer to the "agential" criteria, to which we now turn.
Regarding these agential requirements, for Jaguaribe (1982) they may be one of two types: either to enjoy internal technical-entrepreneurial autonomy, or "a universal intra-imperial relationship" (JAGUARIBE, 1982, p. 23).While the irst requirement, of technical-entrepreneurial autonomy, would involve very high costs for the peripheral countries in a world of growing internationalisation of capitalism and economic interdependence, the second, says Jaguaribe (1982), carries within it the failure of any autonomistleaning project.Jaguaribe (1982) and Puig (1980) also refer explicitly to the strategies for achieving autonomous behaviour, and it is here that agential factors comes into play more clearly.This concerns making alliances with other countries (JAGUARIBE, 1982) and, more speci ically, regional alliances against the centre, in addition to political and economic integration (PUIG, 1980).As Bernal Meza (2013) aptly summarises it, "autonomy was a quality that was built from internal decisions, even when it was systemic conditions that made it possible" (MEZA, 2013, p. 212).
Further developing this point, Puig (1984) goes on to identify degrees of autonomy, from a heterodox position up to potential radicalisation, when it assumes a secessionist character.The irst corresponds to ... the stage in which domestic groups who hold state power, forming part of a bloc, still accept the dominant power's strategic guidance, but openly disagree with it on three major issues: to) the domestic development project, which may or may not coincide with that envisaged by the superpower; (b) international links that are not globally strategic; (c) the dissociation between the national interests of the dominant power and the strategic interest of the bloc (PUIG, 1984, p. 78).
Here the national group in power does not accept dogmatic impositions on behalf of this same bloc, since there are political and strategic considerations corresponding only to the hegemonic power's own interests (MEZA, 2013, p. 215).Radicalisation of this kind of behaviour leads to a secessionist stance, i.e. that in which the state challenges the We must stress that the continuum that Puig (1984) worked with consisted of quarters: paracolonial dependency, national dependency, heterodox autonomy and secessionist autonomy.It is clear, then, that the author has always worked with the opposites of "dependency" and "autonomy".In other words, although he recognises degrees of autonomy, even at its lowest level it is already de ined by opposition -and not by proximity -to dependency.It means that autonomy (regardless of its degree) is always a counterpoint to dependency.This seminal contribution from Puig (1984) should not be minimised.Jaguaribe's (1979Jaguaribe's ( , 1982) and Puig's (1980Puig's ( , 1984) ) contributions were developed during the Cold War context of bipolar competition, and within the framework of a centreperiphery dichotomy marked by strong strategic dependency.However, there is an important dimension in these authors' analyses that has been omitted in subsequent updates and reformulations of the concept in the post-Cold War context: their understanding of the international order as hierarchical3 .As we have seen above, this premise refers to the core-periphery con iguration, and the understanding of the peripheral condition being a de ining feature of these countries' situation.Thus, autonomy cannot be interpreted as sovereignty -a condition that, following Kenneth Waltz (1979), would predominate in an anarchic order in which no power is above the others.In this sense, de ining autonomy as a country's ability to make decisions based on its interest and needs simply echoes Waltz's de inition of sovereignty, and as such it would probably be the dominant strategy of all the states in an anarchic order.But the search for autonomy, according to Puig's (1980Puig's ( , 1984) and Jaguaribe's (1979Jaguaribe's ( , 1982) ) conception, implies overcoming the condition of dependency, not necessarily attaining self-suf iciency or autarchy as it has often been interpreted in the current literature on "the pursuit of autonomy".It is in this sense that Puig (1984) clearly de ines heterodox autonomy as the search for the state's very own model of development, which may or may not coincide with that preferred by the great powers, as we have seen above.
Moreover, for some analysts, emphasising the dimension of sovereignty rather than dependence in reformulations of the concept of autonomy for the contemporary world, makes the quest for autonomy compatible with active participation in international life.In the new order, states try to implement their decisions in an autonomous manner through active participation in global institutions, but this strategy does not necessarily represent an attempt to overcome dependence, as in the original vision of Jaguaribe (1979) and Puig (1980).This reformulation of the original concept was called "relational autonomy", bpsr Letcia Pinheiro & Maria Regina Soares de Lima as we will see below.Ultimately, understanding the autonomist strategy as a pursuit of sovereignty means turning the concept into a parameter common to all states, or a premise relating to the behaviour of states in an anarchic order -what Waltz (1979) called "a selfhelp strategy".As a result, the concept loses any analytical value, as it no longer differentiates between the behaviour of different states.And it certainly loses its normative value as a political objective differentiating the foreign policy strategies of countries situated at the periphery of the global power system and capitalist order.Nothing could be further from the original concept as envisaged by its two founding fathers.
The Argentine academics Russell and Tokatlian (2002) sought to renew the concept in light of the signi icant change to the conditions of possibility for action brought about by the intensi ication of globalisation and the end of the Cold War, at the global scale; and, regionally, by the redemocratisation of many countries and by successful integration initiatives.They proposed a new approach to autonomy, naming it "relational autonomy".De ined as the "capacity and disposition [of the state] to make decisions based on its own needs and goals without external interference or constraints, and to control processes or events that occur beyond its borders" (RUSSELL and TOKATLIAN, 2002, p. 162).As highlighted above though, this de inition does not share the assumptions found in Jaguaribe's (1979) and Puig's (1980) concepts.
Another important point to highlight in Russell and Tokatlian's (2002) proposed model relates to the question of degrees, incorporating the pole of dependency to the scale.In their words, autonomy is ... a condition of the nation-state that enables it to articulate and achieve political goals independently.In accordance with this meaning, autonomy is a property that the nation-state may or may not have along a continuum whose ends are two ideal situations: total dependence or complete autonomy (...).In both cases, autonomy is always a matter of degree that depends fundamentally on the capabilities, hard and soft, of states and the external circumstances that they face (RUSSEL and TOKATLIAN, 2002, p. 162).
Intentionally or not, the authors offer to the analyst a tool likely to generate various interpretations; a mechanism comparable to the metaphor of the glass half-illed with water, which can be seen either as half full or as half empty.Applying this metaphor to our theme, one could equally qualify the behaviour of a state found near the dependency pole as "low autonomy or as high dependency".Despite the signi icance of the adverb -low or high -it is the nouns -autonomy or dependency -that determine the content of the behaviour.As we will see later, this has been a strong tendency in the literature as a result of the politicisation of the concept which, following what has been said above, varies "according to interests and positions of power" (FONSECA JR, 1998, p. 361).
Having made these observations, we now turn to consider how the main contributions to the concept of autonomy have shaped analysis of the guidelines of Brazilian foreign (2018) 12 (3)  e0003 -7/22 bpsr Between Autonomy and Dependency: the Place of Agency in Brazilian Foreign Policy policy from the 1940s to the present day.To put it differently, we aim to explore whether and in what ways the claim that Brazil has adopted an autonomist foreign policy throughout history is, in fact, true according to the original de initions, or whether it is necessary to abandon this de inition in favour of a more appropriate one.

Autonomy, autonomies in Brazilian foreign policy
Throughout history it has been common to portray Brazilian foreign policy as having an autonomist leaning (FONSECA Jr., 1998;MOURA, 1980;PINHEIRO, 2004;SARAIVA, 2014a;SPEKTOR, 2014;VIGEVANI and CEPALUNI, 2011).The examples that we list below clearly illustrate trends in the literature on this subject, showing how the concept has been used in two distinct ways.The irst is contextualising, or "situational", and the second is "behavioural".We will take them as our point of departure to assess to what extent they adhere to the autonomist label, according to Jaguaribe's (1979) and Puig's (1980) original meaning of the concept.
The irst case -which we call a "situational" characterisation -can be exempli ied by the "autonomy in dependency" thesis, a term coined by Gerson Moura (1980).In the early days of the discipline of international relations in the country, Moura (1980) wrote one of the irst papers challenging the structural interpretations then prevalent in the ield, in which Brazilian foreign policy was regarded as innocuous or described only as an epiphenomenon of systemic determinations.Going in the opposite direction, Moura (1980) emphasised the role of Vargas' successful foreign policy in the 1940s.At that time it was possible to take advantage of the Brazilian Northeast's strategic geographical location, and the need for strategic materials for the war industry during the formation of the US power system.This allowed Brazil to attract, among other things, US funding for the construction of the Volta Redonda steel plant, a landmark in Brazilian industrialisation.
Therefore, according to this thesis, even under conditions of strong structural dependence, such as those that prevailed at the time, there were possibilities of agency for peripheral countries, by inding a "way to negotiate the realignment and take advantage of it (...) allow[ing] us to characterise the action of the state as autonomy in dependence" (MOURA, 1980, p. 189).In other words, it would be possible to adopt an autonomistleaning policy in order to circumvent the situation of dependency, thus reaping some bene its.Moura's (1980) add-on to the concept of autonomy -i.e.dependency -was not proposing how it might be achieved in pure form, but pointing to a particular situation in which, despite the adverse conditions, it could be reached.In this sense, Moura (1980) took Jaguaribe's (1979Jaguaribe's ( , 1982) ) proposition further, but without denying it.While the latter saw "national viability and international permissibility as sine qua non" conditions for autonomous behaviour, Moura (1980) placed more emphasis on the importance of agency as the driving force of an autonomist policy in the context of less than ideal conditions, i.e. even in a situation of dependency.At other times, as during the Cold War, foreign policy was an agent of change for the country's international integration.At this time it took advantage of a context favourable to countries of the Geopolitical South, such as decolonisation in the 1960s, and the discussions over possible new arrangements for the international order in the wake of the formation of Organisation of the Petroleum Exporting Countries and the oil crisis of the 1970s.The "independent foreign policy" (1961-64) and the "responsible pragmatism" (1974-79) were characterised by strong autonomist connotations, and both were fruits of a combination between systemic opportunities and the action of agents intent on overturning the country's subordination within the international order.Although foreign policy by itself has not so far been able to change the situation of structural economic vulnerability, it can improve the conditions of the country's international integration, and even diversify its relations of dependence.
It was to this end that the 1970s saw the move to a strategy that the Brazilian diplomat Gelson Fonseca Jr. (1998) has called "autonomy through distance", meaning "a 'quali ied' distance in the debates and negotiations of surrounding key questions during the Cold War period" (FONSECA JR., 1998, p. 360).
It is worth noting that, at the time, the country had made signi icant strides in the condition of "national viability", a criterion that Jaguaribe (1979) and Puig (1980) had indicated as necessary for the exercise of an autonomist foreign policy.During that period, there were positive indicators of economic growth, as a result of the model of industrialisation by import substitution and the investments made in national capacity-building, which expressed elites' commitment to a national project, albeit one that lacked both income distribution and democracy.
We will not go into the strategies used by governments in each one of these two periods -that of the "independent foreign policy" and that of "responsible pragmatism"but is quite clear how their guidelines corresponded to the crucial elements of Jaguaribe's (1979) and Puig's (1980) models: systemic opportunities, national viability and, most importantly, agency.In this sense, the "behavioural" component that Fonseca Jr. (1998) has added to the concept of autonomy -"through distance" -does not contradict its original meaning.In fact, we would underscore how this quali ication shows great similarity to the idea of "heterodox autonomy" elaborated by Puig (1984)  4 .
As the concept of autonomy was no longer understood as situational or contextual 4 For the sake of illustration, it is worth comparing Puig's de inition to Fonseca Jr.'s (1998).According to the former, the degree of autonomy that he labels "heterodox" corresponds, as we have already mentioned, "to the stage in which domestic groups who hold state power, forming part of a bloc, still accept the dominant power's strategic guidance, but openly disagree with it on three major issues: a) the domestic development project, which may or may not coincide with that envisaged by the superpower; b) the international links that are not globally strategic; c) the dissociation between the national interests of the dominant power and the strategic interests of the bloc" (MEZA, 2013, p. 215).For Fonseca Jr. (1998), "the irst expression of 'autonomy' [through distance] would be to keep a distance from the actions of the Western Bloc, especially when they entailed military engagement.We sought to maintain an alignment as regards fundamental values, but we did not turn them into automatic strategic engagement [...]" (FONSECA JR., 1998, p. 362).but rather as behavioural, it began to lose its previous speci icity and, more importantly, to move away from its creators' original meaning.In fact, rather than being an end in itself, as was the case in Puig's (1980) and Jaguaribe's (1979) formulations, autonomy came to be used as a different means of reaching an end, which quite often was not made explicit.
If we look at different cases where analysts have described Brazilian foreign policy as autonomist-leaning, we cannot help but question their adherence to the original concept developed by Jaguaribe (1979Jaguaribe ( , 1982) ) and Puig (1980Puig ( , 1984).Let's see.
"Autonomy through participation" -a term coined by Fonseca Jr. (1998) in the same book in which he also presented the term "autonomy through distance" -refers to the exercise of autonomy through a strategy that promotes adherence to international regimes in order to in luence them, along the same lines as the "relational autonomy" thesis put forward by Russell and Tokatlian (2002).In Fonseca Jr.'s (1998) own words "autonomy translates into 'participation', i.e. a desire to in luence the open agenda with values that express diplomatic tradition and the ability to view the direction of the international order through one's own eyes, and from unique perspectives" (FONSECA JR., 1998, pp. 368-369).
This interpretation would be repeated by several other scholars.To cite just a few examples, Vigevani and Cepaluni (2011) have also used it to describe foreign policy during Fernando Henrique Cardoso's administration (1995-1998 and 1999-2002), describing it as an orientation marked by "...adherence to international regimes, including liberal ones (such as the WTO), without losing the capacity to manage foreign policy.In this case, the goal would be to in luence the very formulation of the principles and rules governing the international system.National goals are thought to be attained more effectively in this way" (VIGEVANI and CEPALUNI, 2011, pp. 35-36).Spektor (2014) changes the label slightly, and uses the category of "belonging" instead of "participation" to explain the years from 1989 to 1999, arguing that during this period "the political and diplomatic leadership did not abandon autonomism.It reinterpreted it in light of new external and internal constraints" (SPEKTOR, 2014, p. 41).
The question we ask ourselves is whether the last two quali iers added to the concept -participation and belonging -would not in fact contradict it.Unlike Gerson Moura's (1980) contribution -"autonomy in dependency" -which gave greater density and complexity to Jaguaribe's (1979) and Puig's (1980) original formulations by demonstrating the possibility of an autonomist behaviour even in a situation of dependency; and equally unlike the add-on suggested by Fonseca Jr. (1998) -"through distance" -which has succeeded in staying true to the main elements of the autonomist thesis, how can one claim now to seek autonomy "through participation" or "through belonging", i.e. through acquiescence?Does this not represent a position of "diminished autonomy", or of "increased dependency"?This is not an easy question to answer and, depending on the perspective we take, it is possible to arrive at different conclusions.Let us take a closer look.Maria Regina Soares de Lima Spektor (2014) claims that during that period particular emphasis was placed on obtaining "good behaviour credentials" (SPEKTOR, 2014, p. 41).This approach is clearly illustrated by the country's accession to the Missile Technology Control regime (1995), by its signing of the Treaty for the Complete Prohibition of Nuclear Tests (1996), by accession to the Treaty on the Non-Proliferation of Nuclear Weapons (1998), by the reduction of trade barriers, the opening of the economy to private investment, etc.But perhaps President Fernando Henrique Cardoso's own words most clearly demonstrate that Brazil had indeed abdicated much of its autonomist behaviour in favour of integration to the mainstream, and that this new behavioural attribute -through "participation" or through "belonging" -strictly speaking nulli ies the autonomist premise: The South is under a double threat -apparently unable to integrate seeking its own interests, and unable to avoid "being integrated" as a servant of the rich economies.(...) In the past it was possible to respond politically to the old relations of dependency by invoking "national autonomy", demanding greater industrial investment to redress the deterioration in the terms of trade, and expanding the domestic market in order to break the chain of the "dependency enclave" and stimulate internal income distribution.Now the political response dictates that the South too should build a new kind of society (CARDOSO, 1996, p. 12).
We said before that the question was dif icult to answer and that it might even elicit different answers.The reason for this is that we must not overlook the investments on regional policy made by Brazil during the same period.Through its rapprochement with Argentina -in resistance to hemispheric integration, and the investments made in the Common Market of the South (Mercado Común del Sur -MERCOSUR) process of subregional integration -the country also tried to establish regional alliances against the centre, i.e. it looked for political and economic integration as an autonomy-seeking strategy (PUIG, 1980(PUIG, , 1984)).
Overall, should we conclude then that these new systemic conditions justify the adaptation of the country's behaviour, and therefore that the add-ons to the concept of autonomy do not contradict it?We do not think this is possible or plausible.If we did this, we would be ignoring the fact that these adaptations and complements go against the original dichotomy that marks the construction of the concept of autonomy as elaborated by Jaguaribe (1979) and Puig (1980), i.e. the ability to distinguish what is or is not autonomist behaviour based on its opposite: dependence.The argument by formulators of foreign policy according to which "adaptations were essential to maintain, not to abandon, the autonomy project" (SPEKTOR, 2014, p. 42) would, in fact, be perfectly congruent with the diplomatic strategy of granting continuity status to the policy through a narrative that always pursues autonomy on the basis of legitimacy, regardless of any epistemological inconsistency 5 .The risk of ending up with interpretations that confuse the analytical argument with the political argument regarding what would constitute an autonomist-leaning policy is further complicated by a very speci ic characteristic of the Brazilian context, i.e. the intimate connection between diplomatic thought and the academic literature on the ield of foreign policy.In these circumstances, the analysis and formulation of the foreign policy get mixed up with one another, with deleterious consequences for the analytical content produced in the ield of Brazilian foreign policy studies 6 .To quote Lorenzini and Doval's warning (2013, p. 16), the justi ication of courses of action through the transformation of what they call "the theory of autonomy" into an ideology, even if it is effective in political and diplomatic terms, certainly does not contribute to advance its heuristic potential -actually, quite the opposite 7 .One example of this situation that, in our view, leads to a weakening of the explanatory potential of the autonomy concept is the belief that, despite "the differences in actions, preferences and beliefs, and having sought very different speci ic results with regard to foreign policy", both Cardoso government and Lula da Silva governments sought "not to distance themselves from the usual quest: to develop the country economically, while at the same time preserving relative political autonomy" (VIGEVANI and CEPALUNI, 2007, p. 275).
The above discussion has sought to demonstrate that the use of the concept of autonomy no longer as an end in itself but as a means has stretched it to such an extent that behaviour like adaptation to changes in the international scene, of bandwagoning with the dominant power, and of acquiescence to the global normative status quo end up being perceived as autonomous because some authors do not determine clearly what the end of foreign policy is.As we have noted earlier, this end may be framed as typical of the selfhelp strategies of countries with limited power resources, situated in the semi-periphery of the international capitalist system, in an anarchic international context.
In their analysis of the strategy of "autonomy through participation", Beleḿ Lopes has proposed a distinction between "foreign policy" and "international politics" which helps us understand this diplomatic narrative.For him, while "foreign policy" meant the diplomatic acquis with its principles, i.e. the so called "permanent element", "international politics" was "a Brazilian norm of conduct within the community of nations, set in face of the problems of the contemporary world, i.e. the mutable element" (CASTRO, 1982, p. 198).Thus, the pursuit of autonomy was "foreign policy", whereas the strategies to reach it were "international policy".In the words of Araújo Castro (1982), "Brazil's international policy has as its primary objective the neutralisation of all the external factors that might contribute to limit its national power" (CASTRO, 1982, p. 212). 6For a discussion of the formation of this ield, see Fonseca Jr. ( 2011) and Pinheiro and Vedoveli, (2012). 7It should be stressed that the interpretations of some past decisions in the context of Brazil's very strong acquiescence to the hegemonic power as indications of an integrationist strategy, and of the latter as a irst step towards an autonomist-leaning policy, have been equally problematic.This seems to be the case of Saraiva's interpretation (2014a) of the inauguration of the Friendship Bridge between Brazil and Paraguay in 1965: "one of the most important historical moves in Brazil's integration with its neighbours.(...) By de initively attaching Paraguay to Brazil, it set the stage for the Itaipu project and for the corporate and economic integration that links both countries and their economies today (...).These examples prove the quick steps that Brazil has taken, in logistic terms, around the dynamic concept of decision-making autonomy" (SARAIVA, 2014a).(2004( , p. 339, apud LOPES, 2013, p. 47) , p. 47) suggest, in line with our argument, that it would be more appropriate to call it "participation by compliance", since it was conceived as an adaptation to the process of economic globalisation that became more pronounced in the post-Cold War era.It was, in fact, a reaf irmation of Brazilian adherence to Western values or, in the words of Chancellor Luis Lampreia, of "adherence to the normativity of the Western canon", as summarised by Lopes (2013, p. 41).Contrary to an autonomist strategy -as this presupposes, an orientation towards revising the status quo -the multilateral activism of Fernando Henrique Cardoso's government might be framed as a typical foreign policy of "prestige", in line with the concept developed by Morgenthau (1971, pp. 67-82).Translated to the present day, a policy of "prestige" aims to use strong multilateral presence as an instrument of soft power to compensate for the hard power not available to it, but without aspiring to challenge the current order in strong terms.
We shall continue to present the periods in which the general line of Brazilian foreign policy has been described as autonomist to the present day, and discuss to what extent how far these analyses adhere to the original concept.
In the 21st century, the "assertive and active" foreign policy of Lula's government -as former Minister of Foreign Relations, Celso Amorim (2015) himself dubbed it -was another period in which the combination of both dimensions, i.e. systemic opportunities and national viability, created conditions for an autonomous foreign policy.It was described by Vigevani and Cepaluni (2011) as a time when foreign policy was marked by the "diversi ication" strategy in its quest for "autonomy".
True to the concept's original de inition, the autonomist label attributed to foreign policy during the Lula years had a "behavioural" component added to it: "through diversiication".This addition is somewhat redundant, since it emphasises a strategy that, strictly speaking, had already been envisaged by Jaguaribe's and Puig's formulation, namely, alliances with other Third World countries (JAGUARIBE, 1982) or, more speci ically, regional alliances against the centre and political and economic integration (PUIG, 1980(PUIG, , 1984)).However, there is an important aspect of Cepaluni and Vigevani's analysis ( 2011) that, although adapted to contemporary circumstances, its with the concept's original meaning -i.e. the realisation that the inclusion of alliances with non-regional partners also serves to increase bargaining power with core countries, something that was epitomised during the Lula period by the IBSA Forum, G-20, G4 and BRICS coalitions.
In the autonomist orientation attributed to the "assertive and active" foreign policy we also ind another trait that Jaguaribe (1979Jaguaribe ( , 1982) ) and Puig (1980Puig ( , 1984) ) had considered indispensable -the formation of a domestic coalition around a model of internal development and a national project which, together, are part of the criterion of "national viability".In this particular case, the coalition formed around a model of state coordination different from the liberal model that predominates in the contemporary order.
In the case of South America speci ically -which saw the creation of the Initia- tive for the Integration of the Regional Infrastructure of South America (IIRSA), launched at the end of the Cardoso government and remaining one of the priorities of the PT (Partido dos Trabalhadores) governments -the differences in interpretation of the PSDB (Partido da Social Democracia Brasileira) and PT governments were striking8 .In the 1990s, both offensive and defensive economic motivations (for meeting the challenges of globalisation) put South America at the center of government foreign policy.Steps were taken to strengthen MERCOSUR, and the IIRSA Initiative was conceived by the end of that decade.The latter envisaged the future of South America as integrated economic area, a market that would expand through the reduction of trade barriers, as President Cardoso stated at the time of its creation in 2000.However, South America was envisaged differently by the Lula government, as became evident right from the start of his government.The new framework for relations with the region revolutionised conventional Brazilian doctrine, emphasising three key dimensions: a conception of integration that included political and social aspects as well as commercial ones; the recognition of structural asymmetries and a willingness to redress them; and the acknowledgement of a strong relationship between Brazil's prosperity and that of its neighbours.These priorities suggest a shift in the collective understandings that had prevailed up until that point, and re lect institutional changes introduced by the PT governments.
These characteristics, which were the hallmark of foreign policy under the Lula government, coupled with the election of the PT candidate to the presidency in the 2011 presidential election, perhaps led us to the over-hasty assumption that in a context where the two structural components needed for the exercise of autonomy were both present, the autonomist orientation would continue without interruption (LIMA, 2015).First were the elements constituting the context of strong "international permissibility", namely: a diffusion of power in the direction of emerging countries; major transformations in South American politics; less overt US presence in the region due to the rede inition of the country's international strategies after September 11; the rise of China; and the emergence of progressive governments of various stripes in South America created very positive expectations.Second was a context of growing "national viability", along with the continuation of the "neodevelopmentalist" domestic political coalition, which led many to believe that the conditions were ripe for Brazil to achieve a less subordinate position in the international system.Maybe this belief spurred voluntarism among the key actors and discouraged any measures to institutionalise the several innovations made during the period, such as: the recon iguration of regional integration, development cooperation, and the establishment of accountability mechanisms for foreign policy.The latter would certainly have strengthened elements of "national viability", thus shielding the country from the period of diminished "international permissibility" that followed.What we have seen, however, is that Dilma Rousseff's irst term triggered the alarm with respect to the continuation of the previous 'assertive and active' foreign policy.This was partly the result of the president's lesser role vis-à-vis her predecessor, in a context where face-to-face presidential diplomacy is increasingly necessary to maintain the status quo, especially for a country like Brazil for whom soft power is one of the main instruments of international projection and action.In addition, the weak institutionalisation of previous initiatives exacerbated the risk of deterioration in what had become a more constrained domestic and international context.
By the end of 2014, the weakening of presidential diplomacy, a worrying economic environment beset by budget cuts, turmoil within the diplomatic corps, open con lict between the Ministry of Foreign Affairs and its young diplomats due to narrowed possibilities for career progression, and waning Chinese demand for commodities already pointed towards a different scenario to that that had prevailed during Lula's two terms of government.The irst few months of Dilma Rousseff's second term saw the deepening of the dif iculties she had faced in her irst due to the worsening economic situation, her political weakness vis-à-vis the Congress, the uni ication of opposition forces and of conservative segments of society, in addition to growing resistance to planned budgetary adjustments by unions and the general population9 .Externally, constraints also increased as global demand declined, commodity prices fell and the Chinese began to prioritise domestic growth to the detriment of exports.Besides this, several of the instruments used in the foreign policy approach of Lula and Amorim, such as subsidised loans from the Brazilian Development Bank (BNDES) to Brazilian investments abroad, and substantial budgetary resources earmarked for international cooperation, suffered a dramatic reduction against the backdrop of recession and iscal adjustment that marked the beginning of Dilma Rousseff's second term.
As a result, might we be watching the collapse of the autonomist orientation of Brazilian foreign policy pursued by the previous government, in light of deteriorating support in Congress and in society, and a less permissive international order?In other words, given the weakening of "international permissibility", the breakup of the domestic political coalition (a central component of "national viability"), and the Rousseff administration's prioritization of domestic issues, can we identify a signi icant change in the design of foreign policy, interrupting the previous autonomist trajectory and shifting towards one of acquiescence, perhaps of dependence?
To answer this question, we will once again make use of the original premises of the concept of autonomy.In doing this we hope to preserve the concept's robustness by rejecting any easy concessions or political resigni ications that distort its meaning.Instead, we favour using it to understand independent actions that, despite perhaps having some bpsr Between Autonomy and Dependency: the Place of Agency in Brazilian Foreign Policy autonomist features, did not manage to establish a paradigm of international integration that might be called a "grand strategy" in Russell and Tokatlian's terms (2015).
Although it is fair to say that the structural conditions of "national viability" and "international permissibility" were present on those occasions when the logic of autonomy guided foreign policy choices, it is also true that they did not guarantee its permanence and continuity.On other occasions, even when the international conditions were permissive, foreign policy choices tended towards the pole of acquiescence and dependence.But we have also seen that even under conditions of strong structural dependency there is enough leeway for the adoption of an autonomist-leaning policy, in order to circumvent this same situation of dependency (MOURA, 1980).From all this we may conclude that the central element in this equation is agency: the nucleus of the concept of autonomy.Without agency there cannot be politics that is critical of the normative structure of the international order.
To recognise the power of structural constraints, however, does not mean thinking of them in a totalising manner.In other words, we suggest abandoning a generalising vision that does not contextualise power or treat its resources as situation-dependent, in favour of one that recognises the possibility of different types of international behaviour resulting from the incentive structure in certain thematic areas, of speci ic power resources in these areas, and of domestic constraints.After all, as we have already said elsewhere, ... a resource that proves to be effective in one thematic area may be irrelevant in another.Thus, capacities and vulnerabilities can vary from one thematic area to another, changing the power relations between these areas.The assumption that power should be measured with respect to speci ic issues leads to questioning the notion of "one single general structure of international power not speci ic to any particular thematic area" (LIMA, 1990, p. 10).
When we recognise the existence of different structures of international power which vary according to the thematic area 10 , and as part of this vision stress the centrality of agency, it is possible to identify positions in the same period that both criticise and contest the normative structure in place.Even though they may not extend to the foreign policy guidelines as a whole, some of these positions may indeed be categorised as autonomist, without having to resort to imprecise gradations or inadequate quali iers in order to label (or to legitimise) the general orientation.Similarly, there is no need to throw out the baby with the bathwater, by giving up altogether the original concept of autonomy.Having said that, let's get back to the Brazilian foreign policy during the Rousseff's government.
Although it would not be accurate to attach the "autonomist" label to this government's general orientation in foreign policy, it would also be wrong to label it "acquiescent".It is not autonomy through "participation", nor "through distance" nor "through diversiication".But as an action paradigm or general guideline, we were also far from joining the mainstream, i.e. adopting a liberal orientation.So while we cannot say that there has 10 For an insight into this analytical perspective, see Lima (1990).
been any deepening or even continuation of the previous general autonomist orientation, neither do we observe a decline in Brazil's international integration (CERVO and LESSA, 2014).Therefore, in assessing this period, we claim that -in line with the requirements of "international permissibility" and "national viability" within the concept of autonomy, and in keeping with a contemporary context in which the variation of capacities and vulnerabilities has been increasing from one thematic area to another -a variety of foreign policy options has emerged, with the result of creating some inconsistencies.Due to an escalation of the political and economic crises during the second term of Dilma Rousseff's government, domestic issues became more relevant within her government's agenda.To cite Celso Amorim, minister of Foreign Affairs of the Lula administration, foreign policy continued to be "assertive" (altiva) but much less "active" (ativa) than during the latter's tenure.
Whichever way we interpret the foreign policy of Dilma Rousseff's government, the president's impeachment has interrupted the cycle of PT governments, which has meant dismantling what had been a foreign policy characterised by its autonomist bias during the Lula/Amorim administration.

Conclusion
Our objective when we began this article was to re lect on the possibilities of ascribing an autonomist behaviour to Brazilian foreign policy, according to how the term was originally formulated.In attempting to answer this question, we have identi ied some gaps, advances and missteps in the use of the concept, as well as some other key observations that we would like to reiterate here.
The irst point to highlight is that the belief in Brazil's supposedly perpetual pursuit of autonomy -the "desire for autonomy", as Fonseca Jr. (1998) has described it -has led analysts to classify as autonomist some behaviours that are antithetical to the original concept of autonomy.
Second, and related to this, is the fact that in Brazil the supposedly distinct activities of formulating and analyzing foreign policy overlap signi icantly.It is perhaps largely due to this situation that the concept of autonomy has been stretched to the point whereby, in practice, it has lost its analytical capacity in accounting for Brazilian foreign policy.In other words, the resigni ication of autonomy was largely the result of a diplomatic strategy that sought to mobilise a narrative of foreign policy continuity.Therefore, the organising principle of Brazilian foreign policy -autonomy -was turned into diplomatic acquis.That is to say that instead of being a standard of conduct that varies according to internal and external context, it is understood as a permanent stance and a political rallying cry (real or imagined), rather than an analytical category used to examine speci ic behaviours.
A third point has to do with the fact that emphasising autonomy as a permanent characteristic and key parameter of foreign policy prevents us from identifying the mo- ments of rupture and even reversals in foreign policy and external relations.We would argue that the 'assertive and active' foreign policy pursued during the PT governments was one of these moments of rupture.
The traditional role of the Ministry of Foreign Affairs in the formulation of Brazilian foreign policy exerted a strong effect of inertia over it, at least until the 1980s and 90s.Moreover the pragmatic-realist orientation that aimed to use foreign policy as a tool of the dominant development model during the postwar period favoured its continuity during this period.Its relative continuity was guided essentially by the ambitions of the Brazilian diplomatic establishment to play a role at the multilateral level, particularly on trade and development issues.This increased international pro ile was interpreted as a quest for autonomy, when in fact it was a typical policy of prestige from a country situated on the periphery of the world system, eager to be recognised by the great powers and to distinguish itself from neighbouring countries and others in a similar position within the global order.
Perhaps one of the most signi icant differences between an autonomist-leaning policy, in line with Jaguaribe's (1979Jaguaribe's ( , 1982) ) and Puig (1980Puig ( , 1984) ) original de initions, and a policy of prestige like the one pursued during the FHC government, is the understanding each had of agency.Those who formulated a foreign policy based on prestige stressed an absence of surplus power and the risks associated with participating in major decisions due to a lack of hard power.In the autonomist view, by contrast, the country should not behave as a "small country" but rather "harden" its soft power, according to the expression used by the then chancellor Celso Amorim (2016, pp. 45-59), although this does not imply becoming a nuclear power.
Due to this belief -that the country should take certain risks on the international scene, given a favourable international and national context -policies with an autonomist bias were moments of rupture and discontinuity in Brazilian foreign policy, every time Brazil adopted a critical position with respect to the status quo.
Moreover, the discussion about autonomist foreign policy directly addresses the structural, systemic interpretations that prevailed in the country in the 1950s, in which foreign policy either had no role or was merely viewed as an epiphenomenon built upon underlying systemic determinations.Although foreign policy agency alone does not have the capacity to alter conditions of structural economic vulnerability, it can modify the terms of the country's insertion in the international order and even diversify its relations of dependence.Emphasizing the role of foreign policy together with a commitment towards a more autonomous development, confer agency on foreign policy as capable of alleviating the unfavorable conditions imposed by the country's peripheral position.Foreign policy implies agency and the possibility of making distinct choices, even in a situation of structural dependency.Autonomist approaches to foreign policy were the result of a combination of systemic opportunities and the actions of agents wishing to alter the terms of the With this in mind, we can identify four periods that were exemplary of this autonomous logic: Vargas's dual policy, the Jânio-Jango governments' "independent foreign policy", the "responsible pragmatism" of the Geisel-Silveira partnership, and the Workers' Party's "assertive and active foreign policy" under the command of Lula da Silva and Celso Amorim.Our discussion of the original meaning of the concept of autonomy leads us to conclude that, contrary to the dominant interpretation of Brazilian foreign policy since the postwar period, which tends to emphasise a more or less continuous autonomist orientation, the prevalence of an autonomist logic in foreign policy was, instead, the exception rather than the rule.

Revised by Matthew Richmond
Submitted on July 24, 2017 Accepted on May 11, 2018 in Brazilian Foreign Policy hegemonic power and chooses to withdraw from the bloc led by it.

bpsr
Between Autonomy and Dependency: the Place of Agency in Brazilian Foreign Policy insertion within the international order.