Do Concepts Matter ? Latin America and South America in the Discourse of Brazilian Foreign Policymakers ∗

By analyzing the content of 6,523 pronouncements by Brazilian decision-makers, this paper aims to present some quantitative evidence to answer the following questions: 01. How did Brazilian Foreign Policy decision-makers employ the concepts of Latin America and South America from1995 to 2014? 02. Were South American terms prioritized over Latin American terms? 03. Did the diplomacy of Luiz Inácio Lula da Silva (Lula) invoke the idea of South America more than that of FernandoHenrique Cardoso (FHC) and Dilma Vana Rousseff (Dilma)? The principal conclusions we have reached are that South America was indeed more cited than Latin America and that this reached its peak during the Lula years. These results are signi icant because they corroborate themain conclusions of historians and qualitative analysts on this matter, through using a different approach based on quantitative methods.

( * ) DOI: http://dx.doi.org/10. 1590/1981-3821201800030006   This publication is registered under a CC-BY Licence.For data replication, see https://doi.org/10.7910/DVN/7ZR0A3Portions of this work are based on the master's dissertation presented by the irst author and supervised by the others at the Federal University of Pernambuco.In this regard, we wish to thank Professors Cristina Carvalho Pacheco and Dalson Britto Figueiredo Filho for their contributions.Any shortcomings are the authors' own.We would also like to acknowledge the inancial support of the Coordination for the Improvement of Higher Level Personnel (CAPES) and of the National Council for Scienti ic and Technological Development (CNPq).bpsr Do Concepts Matter?Latin America and South America in the Discourse of Brazilian Foreign Policymakers H ow did Brazilian Foreign Policy (BFP) decision-makers employ the concepts of Latin America and South America from 1995 to 2014?Was the term South America prioritized over Latin America?Did the diplomacy of Luiz Inácio Lula da Silva (Lula) invoke the idea of South America more than that of Fernando Henrique Cardoso (FHC) and Dilma Vana Rousseff (Dilma)?To answer these questions, the paper presents empirical evidence taken from 6,523 pronouncements 1 by BFP actors 2 .
The purpose of the present work is to advance the body of research on the geographical horizons of BFP.Based on historical analyses, this research agenda has already revealed phases of conceptual transitions in the discourse of key players, such as identity proximity with Europe (1494Europe ( -1902)), America (1902America ( -1961) ) and Latin America (1961-) (COUTO, 2007;SANTOS, 2005;VIZENTINI, 1999).
However, by the beginning of the 1990s, use of terms related to South America grew rapidly in speeches by decision-makers, raising the question of whether or not this had led to diminished emphasis on Latin American terms (COUTO, 2006;RAMOS, 2012;SANTOS, 2011;SPEKTOR, 2010).In effect, this discussion has become a key theme among scholars interested in understanding Brazil's strategies to increase its international in luence, such as Burges (2009;2008;2006), Flemes (2010), and Malamud (2011).Initial research into this issue approached the question through a descriptive, exploratory, and historical perspective (COUTO, 2007; 2006; SANTOS, 2014; 2005 3 ).In addition, recent research has opted to rely on content analysis, as well as on discourse analysis, to reach testable and veri iable conclusions (FERRAZ, 2012;MEUNIER and MEDEIROS, 2013;MIRANDA, 2014;MIRANDA and RIBEIRO, 2015).
Although signi icant progress has been made using such a qualitative and historical approaches, there is still a large gap when it comes to the construction of quantitative indicators that can be used as a proxy to identify minor variances in emphasis on South America vis-à-vis Latin America by actors involved in BFP.Thus, to address this research gap, we assess the longitudinal frequency of Latin America and South America as concepts in the rhetorical repertoire of Brazilian actors, with the purpose of revealing when and how frequently both ideas were invoked.In this strict sense, it is also important to underline that the primary objective of this work is to assess indings in relation to arguments based on data generated through content analysis, and not to provide a historical or theoretical reinterpretation of current views on this issue.Notwithstanding this clear limitation, considering the scarcity of this kind of data in the ield, this paper provides new evidence not 1 Pronouncements, for our purposes, are texts presented in contexts that give meaning to words, allowing the audience of the message to understand what the pronouncer has to say and what is meant when certain words are used. 2Here, we consider as actors all individuals whose pronouncements were included in the database we used as our source, regardless of the in luence they exerted on the decision-making process.As such, diplomats, ambassadors, Chancellors, Presidents, Vice-Presidents, and others are all included.The text is structured as follows: the irst part summarizes both the theoretical discussions and historical contexts that provide the background to the paper, explaining why concepts matter in BFP analysis and, subsequently, offering an overview of scholarly interpretations regarding the operationalization of the concepts of Latin America and South America from 1995 to 2014; the second part presents our methodological approach; the third and inal part presents the main results and raises some questions for future research.

Do concepts matter? Latin America and South America in BFP (1995-2014)
Before presenting our empirical evidence, it is essential to outline the theoretical arguments and historical facts that led us to the research question addressed in this paper.In this regard, we irst explain why concepts matter in BFP and how, by carefully analysing them, we can reach a deeper understanding of the interests guiding Brazilian diplomacy.After that, we examine what the literature says about the use of Latin America versus South America in foreign policy discourse from 1995 to 2014.

Concepts matter
In foreign policy, decision-making processes are multidimensional and depend on a plurality of actors, issues, agendas, visions, and interests (HERMANN and HERMANN, 1989).All of its components challenge the conception of states as akin to billiard balls (HUDSON, 2005).It is insuf icient to consider only correlations of power, systemic constraints, material resources and state capacity.We should also consider the in luence of domestic factors such as leaders, political parties, civil society, public opinion, the executive, legislative and judicial branches, and so forth (FIGUEIRA, 2011;GALLAGHER and ALLEN, 2014;GARRISON et al., 2010;PUTNAM, 1988).Thus, analysing discourses, values, visions, ideas, and roles, as well as conceptual frameworks, is fundamental to better understanding foreign policy (HOUGHTON, 2007;THIES and BREUNING, 2012).
In effect, many foreign policy analysts are turning to constructivist approaches to better incorporate ideational variables into their research designs (WICAKSANA, 2009).Broadly, constructivism can be de ined as "the view that the manner in which the material world shapes and is shaped by human action and interaction depends on dynamic normative and epistemic interpretations of the material world" (ADLER, 1997, p. 322).For constructivists, concepts matter because they are expressions of power, inasmuch as that what prevails is "the notion that ideas are a form of power, that power is more than brute force, and that material and discursive power are related" (HOPF, 1998, p. 177).
Although constructivists adopt similar ontological positions, they differ when it comes to epistemology and the methodological techniques that should be applied to mea-bpsr Do Concepts Matter?Latin America and South America in the Discourse of Brazilian Foreign Policymakers sure and to assess ideational variables (FINNEMORE and SIKKINK, 2001).While there are scholars who design their research based on inferential logic, sustained by traditional quantitative and qualitative techniques, there are also those who criticize scienti ic method as representing a mere product of a hegemonic discourse (ADLER, 1997;KLOTZ and LYNCH, 2007).
In this article, we defend axiological neutrality4 , the strength of arguments constructed from facts and logical reasoning, and the importance of replicability, falsi iability, and validity in research designs (WENDT, 1999).So, for the purposes of this paper, the irst step toward establishing these elements is to clarify what we mean by discourses, conceptual frameworks, and regions in foreign policy analysis (FARIAS and RAMANZINI JR., 2015;KING, KEOHANE, and VERBA, 1994).
Discourses are texts within a given context, where the intention is to materialize and to transmit the cognitive interpretation of a certain agent or institution by means of metaphors, analogies, speech acts, dichotomies, and games of identity and alterity (WICAK-SANA, 2009).Whenever discourses "[involve] diagnoses about national reality, foreign policy priorities and de initions of principles that should guide the country's conduct abroad" (ARBILLA, 2000, p. 340), we consider ourselves to be dealing with conceptual frameworks.This is particularly true for Itamaraty, which appears to take great care when formulating public pronouncements, not only to avoid possible misunderstandings but also to ensure discourse is coherent, based on tradition and continuity with the past (BURGES, 2013;DANESE, 2017).
If the ideas of Latin America and South America are conceptual frameworks developed by Brazilian diplomacy, then it is to be expected that there will be a connection between the frequency with which each term is mentioned and ambitions with regard to regional horizons (COUTO, 2007;WEBER, 1990).It should be noted that geographical signi iers such as regions are based on narratives of unremembered conquests, as well as claims of shared values, identities, and perceptions under the logic of imagined communities (ANDERSON, 2008;SPEKTOR, 2010).In other words, "region is what actors make of it" (NOLTE and COMINI, 2016, p. 550).
Therefore, concepts matter because they reify interpretations of reality.Whenever an actor aims to play a more signi icant role in a certain region, they tend to increase the frequency with which they use terms referring to that geographical unit, both to change meanings and create rhetorical cohesion (OPPERMANN, 2010;WEHNER, 2015).To ind out whether this applies to the cases of Latin America and South America, we must consider both historical and quantitative variation in how they appear in the discourses of national actors.We introduce this question in the next section.The concepts of Latin America and South America in BFP (1995 2014) The questions that we attempt to answer in this paper rest on theoretical arguments that allow us to question whether the concept of South America was used more than the concept of Latin America, and whether or not the salience of the former reached its peak during Lula's administration.
The adoption of Latin America as a geographical label occurred belatedly, not only in Brazil.In fact, because of its exogenous origin, coined in France under Napoleon III, the term only became widely used when the United Nations established the Economic Commission for Latin America and the Caribbean (ECLAC) in 1948 (COUTO, 2007).Furthermore, in Brazilian discourse, its rise was also linked to the decline of the concept of Pan-America, which occurred after setbacks in the special relationship with the United States of America (USA) during the Cold War (MIRANDA, 2014;SANTOS, 2005;VIZENTINI, 1999).
Thus, Latin America gradually became the main conceptual framework used in Brazilian diplomacy, being legally recognized by the Federal Constitution of 19885 .Indeed, prior to this juridical recognition, the idea had already in luenced Brazil's participation in initiatives such as the Latin American Free Trade Association (ALALC), later the Latin American Integration Association (ALADI), and the deepening of regional relations (Galvão 2009) 6 .
From the 1990s onwards, two main factors have hindered innovative initiatives underpinned by the idea of Latin America.The irst is the USA's new positioning with regard to its international insertion, visible in its aspiration to create a Free Trade Area of the Americas (FTAA).This may be seen as a reaction to a bigger adjustment in the international order, whereby the victory of global capitalism was accompanied with movements towards regionalization.As put by Hurrell, in the current global order, one world coexists with many worlds (HURRELL, 2007).
According to Nye Jr. (2004) this new order ceased to resemble the classical unipolar board game that was only focused on military resources, and became much more complex.As he writes: The agenda of world politics has become like a three-dimensional chess game in which one can win only by playing vertically as well as horizontally.On the top board of classic interstate military issues, the United States is indeed the only superpower with global military reach, and it makes sense to speak in traditional terms of unipolarity or hegemony.However, on the middle board of interstate bpsr Do Concepts Matter?Latin America and South America in the Discourse of Brazilian Foreign Policymakers economic issues, the distribution of power is multipolar.The United States cannot obtain the outcomes it wants on trade, antitrust, or inancial regulation issues without the agreement of the European Union, Japan, China, and others.It makes little sense to call this American hegemony.And on the bottom board of transnational issues like terrorism, international crime, climate change, and the spread of infectious diseases, power is widely distributed and chaotically organized among state and nonstate actors (NYE JR., 2004, p. 04).
As a result, the US attempted to bind the American continent to its model of international insertion through the constitution of a hemispheric integration project (TUSSIE, 2010).From Brazil's perspective, this could threaten the country's autonomy, unbalance the national economy, and intensify existing asymmetries (BATISTA JR, 2003).So, Brasıĺia began to articulate coalitions with the States of the region to strengthen its negotiating position, or even avoid the FTAA.Yet, the dif iculty of reaching consensus at the Latin American level increased as Mexico became increasingly interested in the USA's proposals.That was the second factor behind the dismantling of the concept of Latin America, at least as the primary framework of Brazilian diplomatic discourse (SANTOS, 2007).
Although relations between Brazil and Mexico have long ebbed and lowed, they became even more strained from 1994 onwards when Mexico joined the North American Free Trade Agreement (NAFTA) and started to formalize a path of close dependence on its relationship with the USA and the tone surrounding it (LEVY, 2009;MAIHOLD, 2014;SANTOS, 2007).Signing the NAFTA agreement without prior regional consultation caused diplomatic tensions that, according to Rosas (2008), led Brazil to suggest that Mexico should be "expelled from LAIA since it had violated the statutes and the spirit of the institution" (ROSAS, 2008, p. 100).
The more the proposal to create the FTAA and the tensions with Mexico developed, the more Brazilian diplomacy felt uncomfortable with using Latin America as its main conceptual framework (SPEKTOR, 2010).To avoid a potential diplomatic clash, Brazil decided to operationalize a regional imaginary around South America, articulating counterproposals such as the South American Free Trade Area (SAFTA) (SANTOS, 2011).This decision involved a long and delicate process of negotiation inside Itamaraty.In this regard, during an interview in 1997, the former Brazilian Chancellor Celso Amorim described the dif iculties he faced in attempting to cope with bureaucratic inertia and persuade his colleagues that the idea of South America, as conveyed by the initiative of SAFTA, was plausible.As he puts it: For example, I am aware that, operationally, I had to make a lot of effort so that steps could be taken towards SAFTA (...) However, I do not think that it was ideological resistance, but rather resistance due to inertia: "we already have so many worries with Mercosur, why should we start dealing with something else?".We could say it was limitation, more than resistance (AMORIM, 2003, p. 17).
In this very same interview, Amorim (2003)  America as a single political unit and platform to BFP was deeply rooted both on the belief that Mercosur should not be seen as an end in itself and "in recognition of the fact that the current political-economic reality is not so much Latin America, but South America" (AMORIM, 2003, p. 12).For this same reason, he believed that the disappearance of SAFTA only occurred in the realm of public discourse, while the idea that sustained the initiative persisted.
Since the dynamics of this operationalization depend on the political choices made by Brazilian decision makers, it is crucial to offer some observations about FHC, Lula, and Dilma.It should be noted, however, that the aim here is not to provide a detailed discussion of each president, but rather to point out what we should expect to ind empirically from each of them regarding the use of Latin America and South America as conceptual frameworks.
President FHC's irst term (1995 to 1998) can be summarized from both an economic and a political perspective.Economically, the idea of "liberal optimism" prevailed (SILVA, 2008); politically, the notion of autonomy via integration was central (VIGEVANI, OLIVEIRA, and CINTRA, 2003).Both perspectives thus attribute Brazil's interest in joining multilateral institutions to its liberal and democratic character.As a result, decisionmaking processes in foreign policy gradually became more centered on presidential diplomacy and pragmatic institutionalists in Itamaraty7 (BARNABE , 2010;CASON and POWER, 2009;SARAIVA, 2010a).
In terms of conceptual discourse, FHC served as Chancellor of Itamar Franco from 1992 to 1993 and vehemently defended the need to formulate a conception of a distinct South American region (COUTO, 2007;GALVA O, 2009).However, as president, he acted in the opposite direction, causing "a depreciation of the notion of South America in the pronouncements of Brazilian foreign policy" (MIRANDA, 2014, p. 128).Meanwhile, the concept of Latin America continued to be used for geographical purposes and did not undergo a great change in its emphasis, although its meaning only re lected support for hemispheric relations (SANTOS, 2014).
FHC's second term (from 1999 to 2002) was marked by political and economic crises, by the exhaustion of neoliberalism and of the bene its to be gained from the relationship with traditional powers.These factors drove a range of shifts such as a transition from the notion of liberal optimism to a perception of an asymmetric globalization, which led to a deepening of regional relations (OLIVEIRA and TUROLLA, 2003;SILVA, 2008;VISENTINI, 2013).
During this period Brazil proposed one of the main initiatives towards the operationalization of a South America regional platform: the irst South American Summit, held bpsr Do Concepts Matter?Latin America and South America in the Discourse of Brazilian Foreign Policymakers in 2000 in Brasilia (BURGES, 2009;COUTO, 2007;GALVA O, 2009;SPEKTOR, 2010).The meeting prompted signi icant innovations by grouping together leaders of the subcontinent to discuss political arrangements that could "organize a cooperative and coordinated coexistence among states across the sectors of trade, infrastructure, the ight against illicit drugs, and information and technology" (MEUNIER and MEDEIROS, 2013, p. 674).
Here we highlight the material and symbolic results of these innovations.The launch of the Integration of Regional Infrastructure in South America (IIRSA) is one example of a material consequence.As for the symbolic results, the irst South American Summit itself represented the resumption of the idea, initiated by SAFTA but discontinued, towards a new platform built around geographic contiguity as a common feature among members (COUTO, 2007).
The second South American Summit took place in Ecuador in 2002 and saw Brazil reassert its interest in transforming the sub-region into an autonomous unit.This time, between 2000 and 2002, the diplomatic strategy was to promote both a new meaning and salience for the idea of South America, establishing the path that would be followed by subsequent presidents (MIRANDA, 2014;SANTOS, 2013;SCHENONI, 2014;SILVA, 2011;TEIXEIRA, 2011).
Just as FHC altered certain aspects of Brazilian decision-making processes, Lula also established some particularities, such as the shift from pragmatic institutionalism to a more autonomous tendency in Itamaraty, which attempted to develop a discourse based on national development, on the deepening of South-South relations, and on a lofty and active international insertion (FARIA and PARADIS, 2013;SARAIVA, 2010aSARAIVA, , 2010b;;VILELA and NEIVA, 2011).
With regard to the concept of South America, our literature review indicated the systemic emission of this term by Brazilian diplomats.At the same time, it revealed a degree of consensus among authors around the conclusion that the concept became most prevalent in BFP during the Lula years, when it was mentioned in practically all statements (MIRANDA, 2014;MIRANDA and RIBEIRO, 2015;SANTOS, 2005;SARAIVA, 2010b).
In this regard, during Lula's irst term (2003 to 2006)  The second term of Lula's administration (from 2007 to 2010) did not radically alter the prioritization of South America over Latin America.Thus, the signing of the Constituent Treaty of the Union of South American Nations (UNASUR) represented an effort to create an institutional space that could both shape a sense of common identity among members and create a stronger coalition that could potentially shield them during softbalancing9 (FLEMES, 2010;MEUNIER and MEDEIROS, 2013;NOLTE and COMINI, 2016).
Towards the end of Lula's second term, the idea of Latin America was restored by the irst Latin American and Caribbean Summit on Integration and Development (CALC) held in Brazil in 2008, as well as the meeting in 2010 that established the Community of Latin American and Caribbean States (CELAC) through the merging of CALC and Rio Group.Despite this return to a Latin American framing, the focus on South American institutions remained the priority (MIRANDA and RIBEIRO, 2015).
Dilma's irst term as president (from 2011 to 2014) saw activities involving the use of the concepts of both Latin America and South America.To summarise, a vast array of scholars have identi ied a reduction in emphasis and in the assertiveness with which initiatives were proposed and enforced.Furthermore, they claim that presidential diplomacy declined signi icantly, due to domestic turbulence in the Brazilian political scenario and the leader's personal characteristics (CERVO and LESSA, 2014;CORNETET, 2014;PAUTASSO and ADAM, 2014;OLIVEIRA and SILVEIRA, 2015;SARAIVA and GOMES, 2016).
As for the idea of South America, the few existing papers suggest quantitative and qualitative changes.In the case of the former, they identi ied a notable decrease in governmental statements using the term.With regards to the latter, authors consider two main changes: 01.The USA ceased to be regarded as the Other of South American identity; and, 02.The willingness of decision-makers to present Brazil as a regional leader diminished signi icantly, due to the dilemmas and dif iculties of taking on the role of paymaster in a context of growing crisis (SARAIVA and GOMES, 2016).
Our literature review indicated some theoretical expectations regarding the use of the concepts that can be analysed using quantitative evidence.We expect the mentions of South America to be higher in FHC's second term than in his irst.In Lula's years, we expect these mentions to take a quantitative leap and reach their peak.Finally, for Dilma Rousseff's term, we expect the trend to be reversed and to see a substantial fall in the frequency of mentions.
Although as far as we know the authors mentioned here have succeeded in capturing the predominance, at the macro level, of the key concepts, there is still no accurate and disaggregated measurement of how use of these concepts varies from year to year.This is an important task, since it will allow researchers to know, in a detailed and falsi iable way, the trajectory that these concepts follow as they appear and disappear from political bpsr Do Concepts Matter?Latin America and South America in the Discourse of Brazilian Foreign Policymakers and diplomatic discourses.The remainder of the paper presents our effort to map these processes.

Methodology
To preserve the validity, replicability, and transparency of our indings, we summarize10 here our process of data collection and analysis.We detail how we approached the construction of the database and operationalization of variables, explain the methodological techniques we used to quantify qualitative data, and lay out the limitations, assumptions, and opportunities contained within our research design.
We accessed a data source called "Resenha de Polıtica Exterior do Brasil", which is available on the institutional website of Itamaraty11 .The archive contains documents including a myriad of pronouncements from 1995 to 2014, such as essays, interviews, notes, press releases, international agreements, and speeches.
Since they were constructed by the department of communication and documentation at Itamaraty, one of the key aspects of our research design is that the results will be representative of the image of BFP that this institution wants to disseminate.This is something of a double-edged sword.
On the one hand, this is a huge advantage because we can be con ident that the wording of the government's of icial narratives was carefully considered.On the other hand, this also presents a signi icant disadvantage in that it only includes actors, events, and political pronouncements that Itamaraty deemed noteworthy.Metaphorically, we are still con ined by the walls of the Itamaraty Palace12 .
Regarding decision-makers, there are statements from Presidents of the Republic, the Ministry of Foreign Affairs and the Secretary General of Foreign Affairs, as well as some other authorities such ambassadors and vice presidents.We analysed 6,523 speeches from all these decision-makers and present the annual frequency in Graph 01.
It is crucial to emphasize that, one of the main characteristics13 of the distribution presented in Graph 01 is that there are, on average, 326 foreign-policy pronouncements per year; the data dispersion in relation to this mean is very high (standard deviation: 208.3); and the highest values are clustered around Lula's presidency14 .
Although this latter feature might in luence the results obtained, we consider that these asymmetries are just inherited characteristics from the population of speeches analysed.It was the Department of Communication and Documentation of Itamaraty, and not us, that de ined the amount of data collected.
We also emphasize that Vilela and Neiva (2011) same issue and they concluded that this difference is derived not only from the fact that Lula more consistently exercised presidential diplomacy, but also that he had a personal characteristic of being more talkative 15 .In any case, we avoid this potential obstacle by adopting percentage values to control for document length (WEBER, 1990).
Since we are dealing with a speci ic population of speeches, that is, one constructed speci ically by Itamaraty, we decided not to use tools of inferential statistics, so as to avoid unintended generalizations.Thus, the results presented only represent trends identi ied in "Resenha de Polıtica Exterior do Brasil", and not all Brazilian rhetoric on matters of foreign policy.Regarding the operationalization of variables, we attempted to construct a representative, balanced, and parsimonious dictionary (KRIPPENDORFF, 2004;NEUENDORF, 2002).Hence, the concept of South America was symbolized by the frequency of terms such as "South America" and "South American", whereas the concept of Latin America was materialized through the frequency of terms like "Latin America" and "Latin American" 16 .
In addition, the methodological framework was based on content analysis.In this regard, we highlight that "all quantitative models of language are wrong, but some are useful" (GRIMMER and STEWART, 2013, p. 03).Thus, by looking only at numerical frequencies, we are losing critical information, but, at the same time, we can capture an angle not deeply explored by the Brazilian literature, that is, the annual variation of the salience of 15 In addition to these factors, we recognize some intervening variables may have in luenced the distribution of values presented in Graph 01.First, the area of information technology (IT) in Brazil advanced gradually from 1995 to 2014.So, the process of modernization and computerization of Brazilian public agencies in the 1990s was in its infancy.In this context, Itamaraty was not an outlier.Thus, we could expect that the more the institution acquired IT capabilities, the easier it was for the agency to provide data for virtual public access.A second possible set of intervening variables is related to structural constraints and domestic conditions.For example, the frequency of speeches can be a product not only of the personal characteristics of different presidents, but also of economic conditions such as commodity prices, crises, and even varying predispositions of elites to advocate greater or lesser openness to the world. 16Variations of gender, number, and person were included in searches whose languages were Portuguese, English, Spanish, and French.For more information, please consult the methodological appendix.
(2018) 12 (3) e0006 -11/24 bpsr Do Concepts Matter?Latin America and South America in the Discourse of Brazilian Foreign Policymakers the different concepts17 .In this sense, the results presented in the next section build upon, complement and contribute to previous qualitative research.

Results and discussion
Table 01, below, outlines some descriptive parameters of the distribution of terms related to Latin America and South America: Graph 02 presents a longitudinal distribution showing the annual frequency of the appearance of the terms related to Latin America and South America in Brazilian diplomatic discourse.
As can be seen, the concepts followed distinct paths between 1995 and 2014.In general, the trend in the earlier years is for more frequent mentions of Latin America.This changes around 2000, when we see a reversal and mentions of South America become more frequent.
Graph 2. Frequency of use of the terms Latin America and South America per year Source: Authors' elaboration.
Read together, Graph 02 and Table 01 reveal important conceptual aspects that allow us to answer whether South America was, indeed, more cited than Latin America.So, when we look at the mean values, we conclude that whereas the former tended to appear in 1.26% of the speeches, the latter appeared in just 0.84% of them.But these values are bpsr Felipe Ferreira de Oliveira Rocha, Rodrigo Barros de Albuquerque & Marcelo de Almeida Medeiros only representative of the aggregate level.Therefore, we must not neglect the dispersion revealed by the minimum (min) and maximum (max) values, as well as by the high measure of standard deviation for both concepts.
In fact, an intuitive way of looking at this dispersion is by verifying the coef icient of variation.As can be seen, Latin America is more stable than South America.Therefore, to achieve a more accurate answer to the question of which concept is prioritized, we must both to disaggregate the analysis to ind out which was the predominant concept in each year, and assess the degree with which this changed from year to year, as revealed in Graph 03.
We de ined annual dominance as being the difference between the values for Latin America and for South America.There are three scenarios here.The irst is when the result is zero, i.e., the points are touching the black horizontal line.This means that the concepts are mentioned an equal number of times.The second situation is illustrated by positive values, which indicates that Latin America is mentioned more often than South America.The last scenario is when South America is more frequently cited than Latin America, resulting in negative values.
2002 and 2010 were the only years in which there was conceptual equivalence, as shown by the zero values.In terms of conceptual dominance, Graph 03 reveals that South America prevailed in 11 out of the 20 years.Moreover, given that both concepts cover 12.34% of the documents analyzed, it is remarkable that Latin America contributes only 1.93%, whereas South America represents the remaining 10.41% of this value.
In addition, Graph 03 shows that, during the years when Latin America was the dominant concept, the difference in citations was not as great as when South America was dominant.For example, the year that Latin America reached its highest value was 1998, when it was mentioned 0.45% more than the concept of South America.By contrast, in 2005, the year South America reached its maximum value; it was cited 2.46% more fre-( 2018) 12 (3) e0006 -13/24 bpsr Do Concepts Matter?Latin America and South America in the Discourse of Brazilian Foreign Policymakers quently than Latin America.These numbers suggest therefore that the concept of South America was indeed the regional platform prioritized by Brazilian diplomats from 1995 to 2014, at least in their public pronouncements.However, we still must ask whether Lula's diplomacy used the concept of South America more than the diplomacy of FHC and Dilma.To do so, we should irst compare the mean values of both concepts during the different presidential terms, as in Graph 04.
Graph 4. Latin America and South America by presidential terms (mean) Source: Authors' elaboration.
Terms related to Latin America were only mentioned more than those related to South America during FHC's irst term (FHC1).After that, every government prioritized the idea of South America.This raises the interesting question about the trend before FHC1; one which a future research agenda should seek to answer.Graph 04 also con irms the expected results by showing that FHC's second term (FHC2) initiated a rhetorical shift towards South America as the main conceptual framework of regional identi ication.As expected, Lula's presidency increased the frequency of references to South America to its highest point during the period, while Dilma's administration saw a signi icant decline in the frequency of both concepts, although the concept of South America remained dominant.The latter's dominance during Lula's presidency can be clearly seen in Graph 05.
Graph 05 is similar to Graph 03: a positive value means that Latin America was more cited than South America; a negative value indicates that South America was more cited; and a zero value indicates an equal number of citations.Except for FHC1, all presidential terms analysed favoured the concept of South America.Yet, the diplomacy of Lula's irst term (LULA1) can be considered an outlier in the sense that South America was used approximately 1.48% more often than Latin America, a value that is higher than all the other presidential terms together.A future research agenda should address this issue qualitatively to explore this asymmetry between Lula's irst and second terms.
Last but not least, Graph 06 offers evidence for the possible association between the use of the two concepts and key events affecting BFP that occurred during the period.We also employed the measure of concept dominance so that interpretation of Graph 06 is analogous to that of Graphs 03 and 05.We opted to only display events related to the operationalization of South America or Latin America, as discussed in the theoretical section above.These are indicative, not conclusive, results, and we do not propose any causal relationship here.Indeed, we simply wish to highlight a question that should be explored by researchers in future studies.

Graph 6. An indicative association between concepts and events
Source: Authors' elaboration.
The positive trend that had existed since 1995 was interrupted in 2000, when the irst South American Summit occurred.This strengthens theoretical arguments based on qualitative evaluations of the importance of this event for increasing the salience of the concept of South America (COUTO, 2007;GALVA O, 2009;MEUNIER and MEDEIROS, 2013;MIRANDA, 2014;SANTOS, 2005;SARAIVA and GOMES, 2016).After 2000 this trend was reversed and the second South American Summit held in 2002 is associated with a conceptual equivalence in the sense that South America and Latin America were cited with equal frequency, as shown by the zero value.In 2003, Lula came to power and some changes were enacted in Itamaraty, such as the creation of an undersecretary dedicated to South American issues, a shift that inaugurated a trend towards greater use of the concept18 .Thus, in the same year that the position of undersecretary was created the concept of South America was mentioned 1.47% more often than the concept of Latin America.In 2004, with the creation of SACN, South America was cited 0.69% more often than Latin America.And this intensi ied even more with the Summits held by this institution, to the extent that in 2005 South America was mentioned 2.46% more often than Latin America.
2007 and 2008 are signi icant years due to the long preparations for and realisation of two important events: the creation of UNASUR and of CALC, respectively.Whereas the former is an institutional locus that maintains the predominance of South America, the latter represents the return of events at the Latin American level.Indeed, it seems that this dual movement led to a fall in the distance between the concepts, because, for the irst time since 2002, the uses were not as asymmetrical.Lastly, the creation of CELAC also promoted a return to the idea of Latin America, raising the question of whether the focus on South America was merely a project carried out during Lula's presidency or a trend that would continue in BFP over the long term.

Conclusions
This paper has addressed the question of the conceptual frameworks mobilised in the discourses of major actors in Brazil involved in foreign policy issues.We assume that concepts matter because they offer rei ied constructions of social reality.Accordingly, we assessed longitudinal and temporal variation in the frequency with which terms related to Latin America and South America respectively appeared in 6,523 pronouncements available in documentary archive provided by Itamaraty.In particular, we were interested in two aspects of how Brazilian diplomacy used these concepts.
The irst is conceptual dominance, which relates to the question of whether South America was or was not prioritized as a concept.We found that, on average, South America was mentioned with a frequency of 1.26%, while the igure for Latin America was 0.84%.
The second aspect concerns the varying use of these concepts across different presidential terms.We hypothesised that there would be signi icantly more references to South America during Lula's presidency than those of FHC and Dilma.Based on the empirical evidence, we can con irm this hypothesis is accurate.An interesting inding is the asymmetry between Lula's irst and second terms, in the sense that from 2003 to 2006 the concept of South America was mentioned to an unusually high degree, to the extent that the frequency of mentions clearly represents an outlier when compared to the other values.In addition, this paper has alluded to issues which merit greater attention in future research.It is important to study new pronouncements corresponding to the years before 1995 and after 2014 to better understand the long-term trends, continuities and discontinuities in BFP.This is the only way by which we can achieve a more precise understanding of historical variation in the use of Latin America and South America as conceptual frameworks and check whether 2000 was the irst time that the latter outweighed the former or whether this had already happened in the past before 1995.By adding pronouncements after 2014, we can get a better picture of whether South America has remained (and whether it will remain in the future) a conceptual priority, or if it was just a concept overstated by Lula.
Another issue to be approached in future research relates to causation.With this paper we take a irst, relevant step towards establishing it, but there is still a long way to go.Tools like time series analysis will need to be employed to measure the impacts of concepts, new variables will need to be operationalized, and researchers should expand the use of mixed methods in research design.In addition to this, interviews and surveys could be used to better clarify the phenomenon under study.It would also be very useful to employ techniques like Qualitative Comparative Analysis and Process Tracing.
In the sphere of quantitative analysis of texts, many tools have been developed to overcome the limitations of simple frequency analysis and word counts, as were used here.So, Text Mining, Keywords in Contexts (KWIC), measures of words associations, and supervised or unsupervised Machine Learning algorithms are also potentially useful tools for future research.These different research perspectives together form a whole that can reveal how the concept of South America impacted on the rhetorical architecture of Brazil's insertion into the international order.In different ways, they can all bene it from the quantitative approach we have adopted here.We should now continue advancing this research agenda.terms.We ask that researchers carrying out future replications please use the following scheme:
As can be seen, the terms convey the idea of both concepts in both the singular and the plural, and in Portuguese, Spanish, English, and French.They all are between quotation marks as required by the software used in order to count the number of exact expressions.Moreover, we have used the logical Boolean operator 'OR' because we wanted to account for all values related to the concepts.

Documents and OCR
All documents were available in pdf format and have been converted to txt in order for our queries to be made.However, not all of them were editable.We have overcome this obstacle by using the technology, Optical Character Recognition (OCR).We need to report this due to the fact that there are still problems of common misinterpretations of alphanumeric characters for languages other than English.That is also the reason why we have included all txt iles together with the data frame and its codebook.The following chart lists the documents converted to OCR before they were converted to txt.
? Latin America and South America in the Discourse of Brazilian Foreign Policymakers admits that the construction of South 8, several political initiatives(SANTOS, 2005;SARAIVA, 2010b;SPEKTOR, 2010)f this notion, to name a few: 01. the constitution of an undersecretary in Itamaraty concentrated on South America; 02. the third South American Summit, held in 2004; 03. the establishment of the South American Community of Nations (SACN); and 04. the South American Community of Nations Heads of StateSummit, held in Brasilia in 2005(SANTOS, 2005;SARAIVA, 2010b;SPEKTOR, 2010).

Table 1 :
Descriptive parameters of Latin America and South America Matthew Richmond Submitted on February 23, 2018 Accepted on May 22, 2018