Start Submission Become a Reviewer

Reading: Sound and Politics: The Audio Archive of the Brazilian Chamber of Deputies


A- A+
Alt. Display

Research Note

Sound and Politics: The Audio Archive of the Brazilian Chamber of Deputies


Bryan Pitts ,

University of California, Los Angeles, US
X close

Yahn Wagner Pinto,

Duke University, US
X close

Madeleine Roberts

Grosvenor Americas, US
X close


The Câmara dos Deputados, the lower house of Brazil’s National Congress, is a leader in the digitization and online dissemination of audiovisual recordings of legislative sessions. Its website hosts the Arquivo Sonoro, an initiative of the Câmara’s audiovisual department, which contains over 120,000 hours of recordings of speeches, votes, and committee meetings spanning over half a century. The Câmara also makes available online complete video of sessions dating back three years. This research note explores the methodological, historical, and pedagogical opportunities that these audiovisual sources offer scholars and educators. It employs as examples the research of a faculty member, a doctoral candidate, and three undergraduate students, all of whom are among the contributors to a digital humanities project that showcases this remarkable resource. By highlighting ambient noise, volume, emotion, body language, grammatical errors, regional or ethnic accents, and discrepancies with the written record, the authors demonstrate some of the ways that aurality and visuality can enrich and complicate humanities and social science research. They also reveal the multifaceted benefits of institutional support for the digitization of archival material and its free dissemination online to scholars and the public.



A Câmara dos Deputados do Congresso Nacional brasileiro lidera a digitalização e a disseminação online das gravações audiovisuais de sessões legislativas. O seu site sedia o Arquivo Sonoro, que é uma iniciativa da Coordenação de Engenharia de Telecomunicações e Audiovisual da Câmara, e conta com mais de 120.000 horas de gravações de discursos, votos e reuniões das comissões produzidas ao longo dos últimos 50 anos. A Câmara também disponibiliza as gravações completas em vídeo das sessões dos últimos três anos. Esta nota de pesquisa investiga as oportunidades metodológicas, históricas e pedagógicas que estas fontes audiovisuais oferecem aos pesquisadores e educadores. Como exemplos, apresentam-se as pesquisas produzidas por um professor, um doutorando e três estudantes da graduação, que contribuíram em um projeto de humanidades digitais para divulgar este recurso notável. Destacam-se o som do ambiente, o volume, a emoção, a linguagem corporal, os erros gramaticais, os sotaques regionais ou étnicos e as discrepâncias com as transcrições escritas, como formas que a auralidade e a visualidade podem enriquecer e complicar as pesquisas nas ciências humanas e sociais. Também são revelados os múltiplos benefícios do apoio institucional para a digitalização de materiais arquivísticos e sua livre disseminação aos pesquisadores e ao público em geral.

How to Cite: Pitts, B., Pinto, Y. W., & Roberts, M. (2020). Sound and Politics: The Audio Archive of the Brazilian Chamber of Deputies. Latin American Research Review, 55(1), 135–147. DOI:
  Published on 19 Mar 2020
 Accepted on 04 Jan 2019            Submitted on 01 Sep 2017

How can hearing the past change the way scholars understand the people they study? Or, put another way, what can audio sources tell us that textual ones cannot? I first confronted this question in 2012, as I was completing my dissertation in history on the surprising role of the civilian political class in the demise of Brazil’s 1964–1985 military dictatorship (Pitts 2013). As I read transcripts of congressional speeches in which politicians inveighed against the military’s abuses, I encountered the question that lies at the core of the historian’s craft: When could I trust my sources? Newspapers might report that the transcript of a speech criticizing the regime had been edited to avoid offending the generals. Or instead of containing the text of a speech, the transcript might simply state that the speaker had chosen to temporarily withhold it so he or she could “make revisions.” Was there any way to discover what the men and women I studied had really said, I wondered? One day it occurred to me: in all the photographs I had seen, there was one feature that every congressional speech had in common: a microphone. If the speech was being amplified or broadcast, was it also being recorded? On a whim, I emailed the archive of the Câmara dos Deputados, the lower house of Brazil’s Congress. As it turned out, the Câmara had recordings of sessions dating to the 1950s, and the Coordenação de Engenharia de Telecomunicações e Audiovisual (COAUD) had recently finished digitizing and placing online a searchable database of every session since 1970 and every committee meeting since 1996.

It is no exaggeration to say I felt chills when I listened to that first recording, from December 12, 1968. We historians spend our lives trying to coax meaning from sources that are incomplete, contradictory, and biased, that hide as much as they reveal, but seldom can we be sure what words were uttered. As I listened to politicians from forty-four years in the past explain not to me, but to their contemporaries, why they were voting to prevent the generals from prosecuting a colleague who had insulted the military, I did not hear only the words contained in the typed minutes. Nor did I hear a narration directed at me as a historian, as would happen in 2015 when I conducted oral history interviews with some of the few surviving deputies present that day. Instead, I heard the deputies’ conviction as they explained how their honor and sense of duty required them to vote their conscience. I heard their nervousness as they uttered words that not only might (and did) place them on the military’s blacklist, but words they knew would be transcribed and recorded for posterity. I heard their pride, their anger, their jubilation, their fear. I found places where the words they uttered differed in subtle ways from the transcript, because a stenographer had edited it for stylistic reasons, or to lessen the risk of offending the military.

The Arquivo Sonoro of the Câmara dos Deputados offers an opportunity for scholars to explore sound as historical source, even as it raises questions about technical challenges, archival and staff intervention, and political interests that illustrate the danger of privileging sound as a transparent window into the past.1 From 2015 to 2017, I directed the “Sound and Politics” project of the Global Brazil Humanities Lab in the Franklin Humanities Institute at Duke University, with support from the Andrew W. Mellon Foundation.2 The lab offered a space in which faculty, graduate students, and undergraduates from Duke and beyond investigated how this resource can enhance our understanding of Brazilian politics. With support from university digital humanities staff, I designed a website ( to showcase the results of our research into politics, sound, music, and history, and to suggest the possibilities this archive offers for teaching Portuguese, particularly regional Brazilian accents and the differences between the spoken and written tongue.3 This research report, we hope, will encourage other scholars to engage with this archive and similar sources.

The Arquivo Sonoro of the Câmara dos Deputados and Oral History

The digitization of decades of analog recordings, along with the ongoing work of adding to the archive every day Congress is in session, was and is a massive undertaking. But the Câmara dos Deputados and its audiovisual staff benefit from a combination of well-trained public employees, generous funding, and a profound belief in the institution’s importance in Brazilian history.4 In 1996 Silvio de Paula Borges, director of the Câmara department now known as the Coordenação de Engenharia de Telecomunicações e Audiovisual (COAUD), decided to transition from tape to a digital recording format and to begin recording committee meetings in addition to plenary sessions. Over the next several years, the Centro de Informática developed its own indexing software to allow technicians in plenary and committees to index sessions by speaker, in real time. In the meantime, COAUD hired additional staff to digitize acetate records and tapes from as early as 1951 until 1996, using the new indexing software as they went along. All told, as of January 30, 2019, the internal Câmara server hosting the audio files contained an astounding 2.7 terabytes (2,702,363,709,440 bytes) of data, or approximately 108,000 hours of recordings.5 The database is searchable by date, speaker’s name, place given, committee, type of meeting, and keyword. With a few clicks, it is possible to access the audio of every speech that any deputy gave during any date range since 1970. The list of speeches for each session shows the time, down to the second, at which each speech began, and with another few clicks, the site will generate an mp3 file for download of the time range selected.

The Arquivo Sonoro is an excellent example of the potential of the digital humanities, particularly the digitization of archives, as well as the advantages of offering archival collections online, with public access, without restrictions on usage or reproduction. I have been able to find nothing similar for any legislative body in the United States. The audio and video recordings available for the US Congress and state legislatures go back to the late 1990s at the earliest, along with a few recordings held by the US National Archives. Their websites are usually poorly designed and searchable only by date. Brazil’s Câmara should serve as a model for US legislatures and archivists, whose digitization of audiovisual records appears to lag far behind.6

Since the advent of cassette tapes in the 1960s, oral historians have considered what the recorded voice and ambient noise add to an interview.7 By the 1990s, as video recording technology became affordable, they began to consider visual aspects like facial expression, body language, and corporeal movement.8 In recent years, the ability to make interviews publicly available online and to record interviews and informal interactions on a smartphone have come to the fore.9 What sets the Arquivo Sonoro apart is the person toward whom the voice is directed. Oral history is an inescapably retrospective endeavor, a performance that entangles memories of the past, the interviewer’s and interviewee’s interaction in the present, and the goals that both want the interview to serve in the future.10 In the Arquivo Sonoro there is no interviewer. Instead, the voices are directed at a host of other actual and anticipated interlocutors: the speaker, deputies watching from their desks; the galleries; the press with their voice recorders and cameras; the stenographers; constituents who might eventually read, listen to, or watch the speech; and, perhaps, the future citizen or historian who might listen to the speech and judge its speaker accordingly. The Arquivo Sonoro is pervaded by a particular sense of immediacy.

Another difference between oral history and the Arquivo Sonoro lies in the people whose voices they preserve. Oral history, whose origins lie in the “new social history” of the 1960s and 1970s, often strives to give voice to those silenced by the official (documentary) transcript—workers, women, people of color, LGBTQ+ people. The people who speak through the Arquivo Sonoro have never been voiceless. Their words—nearly always those of wealthy, white men—have already been recorded in newspapers and transcribed in congressional records, and now the Arquivo Sonoro meticulously preserves and makes their words available to a global audience, even when they speak on mundane topics or engage in schoolboy brawls. The fact that such efforts are taken to preserve, catalog, index, digitize, and disseminate these voices (but not others) speaks to the priorities of Brazilian political institutions. It renders even louder the voices of those who have always had the power to make themselves heard.

Sound, the Arquivo Sonoro, and the Construction of History

What are some ways in which scholars might use audio sources, and specifically the Arquivo Sonoro of the Câmara dos Deputados? The most obvious use is to obtain information that the written record leaves out. In the case of the Câmara dos Deputados, the daily transcript often omits parts of a session that are not sanctioned by the Câmara’s internal rules. For example, in 1992, when the lower house voted to impeach President Fernando Collor de Mello, when the roll call begins, the transcript simply states “Processo de Votação.” However, the only words transcribed are two procedural questions. A roll call of over two hours, in which 480 deputies voted, is condensed into a few sanitized lines (Diário da Câmara dos Deputados, September 30, 1992, 120–122).11 The audio record reveals that about three quarters of the deputies did not simply respond “Yes” or “No,” but added brief declarations explaining their votes. That is, the audio shows how politicians justified their actions at a moment when they knew that the eyes of Brazil were upon them, one of the few moments in their careers when they could be certain that they were making history. The people and political causes they dedicated their votes to thus offer significant insights into the political culture of Brazil’s elite. “For the factory workers, for the port workers and stevedores, I say yes.” (Pelos operários, pelos trabalhadores portuários e estivadores, eu digo sim.) “For the evangelicals of the state of Rio de Janeiro, and for morality, yes.” (Pelo povo evangélico do Estado do Rio de Janeiro, e pela moralidade, sim.) “For the civil servants who have been so disrespected, for landless workers, for the Brazilian people, for the street children, without fear of being happy, out with Collor.” (Pelos servidores públicos tão desrespeitados, pelos sem terra, pelo povo brasileiro, pela meninada da rua, sem medo de ser feliz, fora Collor.) (Arquivo Sonoro September 29, 1992, 15:59:15).12 More intriguing still, it is possible to place these forgotten statements alongside the far more publicized ones made during the 2016 vote on the impeachment of Dilma Rousseff to discover how the most common themes of 1992 are similar to and different from 2016.13

Media File 1 

Deputies justify their votes to impeach Fernando Collor. Arquivo Sonoro, September 29, 1992.

Yet parliamentary rules are not the only reason information might be left out of the transcript. Particularly during the military dictatorship, the often-tense relationship between Congress and the military kept congressional leaders in a state of vigilance, lest a rogue legislator offend the military and provoke an overreaction (Pitts 2011). This is what happened on April 28, 1980, when Deputy João Cunha blasted the military for repressing striking metalworkers in São Paulo. Yet the transcript says, “Mr. João Cunha gives a speech, withheld for revision [O Sr. João Cunha pronuncia discurso, retirado para revisão]” (Diário da Câmara dos Deputados, April 29, 1980, 2,779). The press, recently freed from censorship, was not eager to spark a crisis and did not report the content of the speech. The only place it was preserved was in the audio of the session. “Yoked to corruption, strangled by hidden ties, controlled by the powerful, they have no explanations beyond lies, violence, and explosions of authoritarianism and the clownish spectacle of the half-dozen generals who sustain them.” (Atrelados à corrupção, garroteados pelos compromissos escusos, controlados pelos poderosos, não têm outras explicações se não a mentira, a violência, explosões de autoritarismo e o espetáculo à palhaçada de meia dúzia de generais que os sustentam.) Cunha read quickly as he tried to fit his speech within the five-minute limit (Arquivo Sonoro, April 28, 1980, 14:04:36). This may have been the harshest speech ever delivered in Congress against the regime. It sparked a panic at a moment when the generals’ process of liberalization (abertura) was still fragile. And if not for the audio recording, it would have been lost.

Media File 2 

João Cunha criticizes the military regime. Arquivo Sonoro, April 28, 1980.

Another reason the transcript and audio might not match is related to the formal and informal conventions of the Portuguese language. While Congress remains, on average, far better educated than the Brazilian populace at large, there are quite a few deputies from the poorer states of the North and Northeast or from rural areas, and not all of them are university graduates, or even secondary school ones. Their chief qualification may be not their political experience but their status as the relative or close associate of a local boss (coronel). Even in the more prosperous South and Southeast, the broadening of the field of popular participation that accompanied the fall of the military dictatorship meant that Congress has seen deputies who were previously activists in favela associations, soccer players, and even a circus clown. And, although it says nothing about their legitimacy as representatives of their constituents, the fact remains that many of these deputies speak nonstandard Portuguese. Their language is that of the factory, the farm, and the street, not the courtroom or classroom. Indeed, perhaps one of the most important characteristics of Brazilian Portuguese is the vast distance between the culto (cultured) language of literature, legal briefs, and (for decades) congressional speeches and the grammatically simplified, often mispronounced Portuguese of the working classes.

When they arrive in Brasília, deputies with less formal education are unprepared for some of the unspoken (but no less rigid) codes that now rule their conduct. And the army of career civil servants—aides, archivists, stenographers, legal advisors, and others—who are the backbone of Congress tend to be educated members of the upper middle class who have their own ideas of how a deputy should act, speak, and sound. Thus, the stenographers who produce the notes that serve as the basis of the transcript are granted significant latitude to correct deputies’ grammar. Of course, as a former stenographer told me in a 2015 interview, many deputies appreciate this; eager to disseminate copies of their speeches to their constituents back home, they want it to be in “good” Portuguese.14 The stenographers help them appear more educated than they in fact are. The result, however, is that class differences among legislators can be blurred. Take, for instance, a 1987 speech by the newly elected Luiz Inácio Lula da Silva, founder of the Workers’ Party and future two-term president of Brazil. Lula was a career metalworker and unionist, the son of impoverished Northeastern migrants, who had never studied past fourth grade, and his spoken Portuguese reflected that. In the first speech he gave as a deputy, the differences between the transcript (left) and his actual words (right) are striking.

Anuncio este evento na medida em que se percebe não haver destaque nenhum na imprensa, embora seja de muito maior importância do que a presença dos cinco fiscais do FMI que estão no Brasil fiscalizando as nossas indústrias, nossas contas, conforme diariamente a imprensa destaca (Diário da Assembléia Nacional Constituinte, May 14, 1987, 1936). Eu gostaria de aproveitar para anunciar esse evento na medida em que a gente percebe que não tem tido nenhum destaque na imprensa, embora seja um evento muito maior do que a presença das pessoas do FMI para fiscalizar as contas brasileiras. Cinco fiscais do FMI estão no Brasil fiscalizando as nossas indústrias e as nossas contas, e a imprensa todo dia dá destaque (Arquivo Sonoro, May 13, 1987, 16:43:10).

Media File 3 

Lula announces a labor congress. Arquivo Sonoro, May 13, 1987.

In the actual recording, Lula uses the nonstandard (but widely common) a gente percebe in place of the first person plural conjugation (percebemos); the transcript changes it to the passive voice, se percebe. The informal construction a imprensa todo dia dá destaque (the press places a highlight on this every day) becomes conforme diariamente a imprensa destaca (in accordance with what the press highlights daily). These and other changes throughout Lula’s single term as federal deputy do not change his speeches’ meaning, but they do present a Lula who is rather sanitized, who speaks like a politician is supposed to speak (albeit without the rhetorical flourishes of his more educated colleagues). The Lula who worked on the shop floor, who went out for cachaça with his friends after their shift ended, whose earthy language and ability to relate to the masses would later make him the most popular and beloved president Brazil has ever known—this is the Lula who begins to shine through in the audio.

The preceding examples show an obvious use of audio: to fill in the gaps left by the written record. In this sense, the audio functions much like the written word, as a transcript of the deputies’ words. However, there is also a less obvious but more methodologically innovative use for audio recordings: to transmit certain aspects of communication that words cannot. Take for example the speech that Ulysses Guimarães, leader of the largest opposition party during the military dictatorship, gave on April 24, 1984, the eve of a vote on a constitutional amendment to restore direct presidential elections, suppressed by the military nearly two decades before. The vote followed five months of mass demonstrations urging its passage, the largest political protests Brazil had ever seen. Guimarães had been educated at the country’s most prestigious law school and was a master of the Portuguese language, his erudite speeches filled with historical and literary allusions. All that is apparent in the transcript as he recalls how inspired he was to see millions filling the streets to demand the right to vote for president. “I saw the people be born of the masses. I saw the rainbow radiating the alliance between the workers and democracy, I saw the disgraced, the dispossessed, and the unemployed convince themselves that there are no rights or wellbeing without citizenship, and that if bad politics destroy them, only good politics can save them.” (Vi o povo nascer da massa. Vi raiar a aliança entre os trabalhadores e a democracia. Vi os desgraçados, os despossuídos, os desempregados convencerem-se de que não há direitos nem bem-estar sem cidadania, e que se uma má política os destrói, só uma boa política pode salvá-los) (Diário da Câmara dos Deputados, April 25, 1984, 2,407). But the audio reveals that not only did he believe that an occasion of historical import required a speech that was beautifully written, but also that it needed to be beautifully delivered. Today his sonorous voice, echoing through a nearly silent chamber, seems exaggerated, even pompous. His sentences begin with the high pitch of a prophet proclaiming the coming of a savior (in this case, the direct vote); they end far lower, shifting the emphasis from the joy of approaching democracy to the weight of the responsibility that has fallen on the deputies as they make this historic decision (Arquivo Sonoro, April 24, 1984, 15:28:27). Especially when compared with the far more informal and less carefully delivered speeches that prevail in the Brazilian Congress today, Guimarães’s speech is a relic of a bygone era, one in which congressional and party leadership was dominated by educated scions of elite families more concerned with a speech’s rhetorical beauty than its popular accessibility.

Media File 4 

Ulysses Guimarães defends direct elections. Arquivo Sonoro, April 24, 1984.

A single recording might contain examples of several of these characteristics. On October 26, 1965, the Senate and Câmara met jointly to debate a series of constitutional amendments that would increase the power of the military-dominated executive branch at the expense of the legislative branch. Senator Artur Virgílio gave an impassioned speech in which he accused regime-allied politicians of cowardice for voting to reduce their own power. “A Congress that is besieged, a Congress suffering pressure like it never imagined it could suffer, a Congress without autonomy that tomorrow could find itself beneath another Institutional Act and have several of its members purged from office, this is a Congress who refuses to humiliate itself? This Congress that has relinquished everything under pressure.” (Um Congresso sitiado, um Congresso sofrendo pressões como nunca imaginou pudesse sofrer; um Congresso sem autonomia que amanhã poderá estar sob outro Ato Institucional, e ter vários dos seus membros cassados, este Congresso é que não se agacha? Este Congresso que tem cedido tudo sob pressão.) When Virgílio alludes to the possibility of legislators being cassados (removed from office with the suspension of the right to run for office or vote for a decade), he practically shouts; the anger and desperation in his voice are clear, certainly because he knew that his own career would be in jeopardy. (In fact, Virgílio was later removed from Congress by the military, not in 1965, but in 1969, under the infamous Ato Institucional No. 5.) When he accuses his colleagues of having “relinquished everything under pressure,” a chorus of shouts comes, “I protest! I protest! [Não apoiado! Não apoiado!]” This reaction, permitted by congressional rules, the transcript includes. At this point, the transcript simply says, “Tumult in the Chamber. The speaker rings the bell” [Tumulto no Plenário. O Sr. Presidente faz soar as campainhas],” a bit of an understatement, since this shouting match lasts a minute and a half, as the deputies hurl insults back and forth while the speaker cuts their microphones, rings the disciplinary bell, and shouts in vain for calm (Diário do Congresso Nacional, October 28, 1965, 754; Arquivo Sonoro, October 26, 1965, 21:49:33).

Media File 5 

Artur Virgílio provokes a tumult. Arquivo Sonoro, October 26, 1965.

The above examples represent some of the uses I have found for audio sources in my own research. Yet as I have seen students work with the Arquivo Sonoro, I have been humbled as I realize the limitations my own interests and discipline impose upon me. Students have made their own discoveries in the Arquivo Sonoro and in more recent video footage of sessions. In the following two sections, two Duke University students, one a Brazilian doctoral candidate in Music and the other a 2017 BA in History from the United States, explain in their own words the uses to which they have put audio and video recordings. I also discuss a collaborative project of my students in a 2017 course on Brazilian politics at the University of Georgia, in which, using video footage, each student analyzed the audiovisual aspects of several deputies’ votes during the April 2016 opening of impeachment proceedings against President Dilma Rousseff.

Sound and Brazil’s First Indigenous Congressman

The following section was written by Yahn Wagner Pinto, composer and PhD candidate in Composition, Duke University.

As a composer, my main interest in the recorded speeches relates to sound and the various ways it produces meaning. Sound is a product of bodies, and in this case, the body of a specific federal deputy, Mário Juruna. Born in 1942, Juruna served in Congress from 1983 to 1986. A member of the Xavante ethnic group, he was the first indigenous member of Congress. Juruna appeared on the national stage as an indigenous leader in the 1970s, when he gained attention for his arguments with government officials about indigenous land rights. To prove to communities back home that federal officials were lying, he used to carry a recorder to tape his conversations with them. The recorder was part of his body as an indigenous leader, but it was the microphone that became the essential part of his body as a politician.

The Xavante were contacted by the Brazilian government only in the twentieth century, so Juruna did not learn Portuguese until he was an adult. As a result, his speeches were full of unique characteristics, ranging from lexical transformations to intonation patterns. Intonation is the linguistic term that describes the variation in pitch our voice produces when we are speaking, the melody of our spoken voices. Juruna’s intonation patterns were much more expressive than those of other politicians, with a pitch range twice as large as other prominent legislators. Specifically, I compared his pitch range with those of Senator Fernando Henrique Cardoso and Deputy Fernando Collor de Mello, both of whom were later elected president of Brazil and who represent the traditional elite who had always been Brazil’s political voice. The official transcripts were unable to express the very clear intonation contour properly registered in the recordings.

Another important aspect of his speeches was the consistent ways Juruna transformed some words. For example, he always pronounced the verb respeitar as “simirespeitar,” probably because Brazilians often use the expression vê se me respeita (Arquivo Sonoro April 19, 1983, 14:35:13). Juruna may have understood it as “vê simirespeita,” and then brought that word to his lexicon in that form. The official transcripts edited many of his unique lexical transformations, trying to fit his speech into a more standard Portuguese (Diário do Congresso Nacional, April 20, 1983, 2,037). However, the press emphasized his lexical deviations, as a purposeful project to discredit him politically. In addition, the recordings captured the ambient sound during Juruna’s speeches, with shouts and applause, showing how his speech affected his audience on an emotional level.

After analyzing these features of Juruna’s speech, I decided to write a piece of music in which I would explore the original recordings, mixing them with instrumental sounds (Wagner 2017). I decided to choose three different moments of his life: 1983, the first year of his term; 1986, his final year in office; and 2002, the last year of his life. For 1983 and 1986, I chose speeches in which he addressed issues of importance to indigenous people and criticized the two largest political parties of the time, the PMDB (Partido do Movimento Democrático Brasileiro) and PFL (Partido da Frente Liberal) (Arquivo Sonoro, April 19, 1983, 14:35:13; April 9, 1986, 23:57:38). For 2002, I used a recording of the VII National Human Rights Conference, organized by the Câmara’s Human Rights Committee, at which Juruna spoke, just one month before his death from diabetes complications (Arquivo Sonoro May 15, 2002, 17:03:13). His voice had become dramatically different. He had a much lower-pitched tone and had difficulty breathing. My composition emphasizes the expressive nature of his speech, even in his last days of life, an aspect that was erased from the official transcription but which was his greatest political asset.

Media File 6 

Mário Juruna as a federal deputy. Arquivo Sonoro, April 19, 1983.

Media File 7 

Mário Juruna before the Human Rights Committee. Arquivo Sonoro, May 15, 2002.

Media File 8 

Yahn Wagner Pinto, Magina: In Memoriam Mário Juruna, para septeto e eletrônica.

Juruna had a voice, and his voice expressed the way he lived: the oppression, the struggle, the friends and family who died because of misguided or deliberately hostile indigenous policies. Everything was there, in his voice. He knew he could connect with people using his voice, even though he was not using a very standard form of the Portuguese language. This self-knowledge is something that transcriptions cannot express but is powerfully evident in audio recordings: an essential aspect for understanding his political personality.

Sound, Sight, and Debates over Domestic Workers

The following section was written by Madeleine Roberts, BA in History/Public Policy Studies, Duke University.

My 2016 senior honors thesis in History examined the roles of domestic workers in Brazil (Roberts 2017). In the second chapter, I analyze the 2013 constitutional amendment that granted a host of labor rights to domestic workers, the only workers’ contingent to have been excluded from such protections since the 1988 Constitution. At the helm of the fight to pass the amendment was Benedita da Silva. A former domestic worker herself, Silva was the first African-descended, working-class woman to ascend to high political office—as Rio’s governor, a senator, and now a federal deputy. A year after the constitutional amendment passed, she gave a speech in Congress in her old maid’s uniform (fardinha), replete with a starched white lace collar and apron over a navy dress, to commemorate the National Day of the Domestic Worker.15

I would never have been able to recognize the effort behind and importance of this legislation and its aftermath without the audio and video recordings from the Câmara. Da Silva’s manifold statements on the matter were inspiring—she has a way with words. In the video recording of her 2014 speech commemorating the National Day of the Domestic Worker, I could notice the subtle detail of her rubbing her forearm while saying that the fight for domestic workers’ rights was one of “the skin” (da pele), implying the racial disparities of domestic employment relationships. In stark contrast was another Rio de Janeiro deputy, Jair Bolsonaro, a divisive figure from Brazil’s far right who in 2018 was elected president of Brazil. The most outspoken opponent of the amendment, one week after the Chamber approved the legislation he stated, “It is absurd [that]: if my nanny, for example, has a child up to 6 years of age, I have to pay daycare for my son’s nanny’s son” (Arquivo Sonoro November 21, 2012, 15:28:39).

Throughout my thesis, I incorporated as much of this detail as possible. My second chapter leaned on the legislative material heavily, and the videos and audio recordings gave my chapter a much stronger narrative. The constitutional debate illuminated long-standing political rivalries, class divisions, and the political symbolism of domestic work. The video and audio recordings of the debates enabled me to depict that in full color.

Sound, Sight, and the Impeachment of Dilma Rousseff

The following section was written by Bryan Pitts, with quotations from audio/video analyses written by two University of Georgia undergraduates.

In Spring 2017, I taught a course titled “Brazilian Politics and Political Culture” in the Department of Romance Languages at the University of Georgia. Although the course was mostly taught in English, the twelve undergraduate students all spoke Portuguese, including three students in the university’s federally funded Portuguese Flagship who had completed a year in Brazil, and two children of Brazilian immigrants. One of their assignments was to select a Brazilian state and analyze the audiovisual aspects of three of its deputies’ votes in the April 17, 2016, session in which the Câmara recommended the impeachment of Dilma Rousseff. I asked them to focus not on what the deputies said but on how they said it and how their colleagues responded—things unavailable in the written transcript. The visual aspect was possible because the Câmara began archiving video of its sessions online in 2012. As the students demonstrated, the richness of analysis is deepened not only by hearing history being made but by watching it unfold.16

Lilian Zhu ambitiously chose to study São Paulo, the state with the largest delegation. She demonstrated a sharp eye for the body language of the deputies crowded around the floor microphones as their colleagues gave brief speeches explaining their votes. In her analysis of Deputy Ana Perugini, of Rousseff’s Workers’ Party (PT), she pays special attention to gender.

When delivering her short speech, Perugini seems to be a little more anxious than the rest: she stumbles on the word “youth” [juventude] and looks around nervously for the first half. She also appears to be speaking very loudly, and for good reason: even though the crowd of deputies is fairly restless during the entire impeachment proceeding, they are particularly disruptive during hers. All the men behind her are turned in the opposite direction, blatantly uninterested in her testimony. Two are laughing about something with a third party who stands outside the camera frame; another one blows a kiss to someone at the front of the chamber. Perugini is already at a disadvantage as a “No” vote from the PT, and the indifference that other deputies show during her testimony also raises the issue of sexism in Brazilian politics (Zhu 2017).

Media File 9 

Ana Perugini impeachment vote. Câmara dos Deputados, Sessão Deliberativa Extraordinária, April 17, 2016.

Daniel Cleveland analyzed deputies from Pará. The attention to detail he shows as he focuses on Deputy Wladimir Costa’s cadence and volume is particularly impressive.

Wladimir Costa [speaks] like a preacher, with a booming voice and great conviction. He refers to all the people that he represents in … Pará, like the people in meu Sudeste, his amigos in the Solidariedade Party, and his children, with a pause in between each to put emphasis on the fact that he is voting with them in mind. At this point Costa is practically yelling into the microphone. He goes on to say that his vote is sim for impeachment, that, “We vote yes!” [Nós votamos sim!] Each yes [sim] is long and drawn out, emphasizing how passionately he supports impeachment. He speaks with such conviction that it seems like he truly believes that he is voting the way everyone in Pará would vote. To put a final point on his vote, Costa shoots off a firework into the crowd. Costa never cites a real reason for his vote …; rather, he just mentions all the people he is voting for. Still, his vote was an entertaining display and was met with cheers from the crowd (Cleveland 2017).

Media File 10 

Wladimir Costa impeachment vote. Câmara dos Deputados, Sessão Deliberativa Extraordinária, April 17, 2016.

Both students focused on details that were indiscernible in the transcript. With all the noise in the Câmara that day, the crowd of deputies pushing against one another, the sense of dread for some and anticipation for others, how far can words go toward helping us understand the events of that day, perhaps the most important in Brazil since April 1, 1964, when Congress endorsed the military’s coup against leftist president João Goulart? Body language, facial expressions, tone of voice, cadence, and reactions from others tell us how that day felt in a way that words on a page never can. And the ability to see what was happening and not simply read or even hear it adds layers of insight that would be impossible otherwise.


That day in 2012 when I first listened to the Arquivo Sonoro, I hoped that sound might solve some of the conundrums presented by textual sources. The reason I had requested the recording from December 12, 1968, was because there were multiple written versions of a famous speech by Deputy Márcio Moreira Alves, as the Câmara prepared to vote on whether to strip him of his parliamentary immunity so that he could be prosecuted for offending the military. Perhaps the recording would allow me to finally know what had “really happened.” When Moreira Alves began to speak slowly and measuredly in his gravelly baritone, I hung onto every word. “Mr. Speaker, fellow deputies, fate has chosen me to be transformed into a symbol of the most essential of the prerogatives of the legislative branch.” (Senhor Presidente, senhores deputados, marcou-me o acaso para que me transformasse em símbolo da mais essencial das prerrogativas do Poder Legislativo.) And then silence, broken only by the scratch of the acetate tape, that stretched on and on before giving way to another speech, probably one that the December 12 session was recording over. (In the 1960s, before the Câmara established a standard procedure for recording, sessions were often recorded over.) Most likely, the Speaker, fearful that Moreira Alves would say something else to offend the military, instructed the technical staff to stop recording when he started speaking.

At that moment, I was cured of any notion that sound might provide access to an unmediated past. The silence made it clear that someone had believed that Moreira Alves’s speech was too dangerous to record. The half-recorded-over speech that came after showed that some speeches were judged unworthy of preservation. And for whatever reason, this partial session has never been placed in the Arquivo Sonoro; I obtained it as an email attachment after making a specific request for that session. There are many other times in the 1960s when partial transcripts are included in the online archive, but this day was left out. Why? Like written sources, audio sources are constructed. They are mediated. They are conditioned by their culturally and historically informed contexts. They exalt some voices and silence others. They respond to the needs of the political moment, not those of the future scholar.

Just like written sources, however, the problems with audio sources do not mean that scholars cannot use them, for their silences and biases add layers of meaning that enrich our understanding of the past and present. As research in Duke’s Global Brazil Lab and Portuguese courses at the University of Georgia have shown, aurality (and visuality) are helping scholars and students develop fresh ways to approach history, politics, and communication. In addition, surely other Latin American institutions have seen fit to store recordings of their legislative sessions, board meetings, union assemblies, or media advertisements. Yet what is being done to preserve these recordings, many of which are likely stored on fragile records, tapes, or reels? And how can scholars and archivists ensure that they are not only preserved, but made publicly available? The Arquivo Sonoro offers examples of how this can be achieved.


1The Arquivo Sonoro is available at 

2The Global Brazil Lab was directed by John D. French (History), Esther Gabara (Romance Studies), Paul Baker (Earth and Ocean Sciences, 2014–2015), and Christine Folch (Anthropology, 2015–2017). This research grew out of innumerable conversations with John D. French, and the idea of developing a collaborative project based on my use of the Arquivo Sonoro in my dissertation was in large measure his. 

3Christina Chia (Franklin Humanities Institute) and Brian Norberg (Trinity Technology Services) at Duke, and Emily McGinn (UGA Libraries Digital Humanities Lab) at the University of Georgia provided invaluable assistance as I set up and designed the project website. 

4The Arquivo Sonoro exemplifies a broader trend in the Brazilian Congress toward developing more comprehensive and transparent communication channels with the public. A recent analysis of TV Câmara and Agência Câmara (the Câmara dos Deputados’ television network and online news service, respectively) concluded that they “contribute to enhancing the accountability and quality of Brazilian democracy, insofar as they provide access to relevant information about legislative processes” (Lemos, Barros, and Bernardes 2016, 201). 

5Information provided by Cînthia Neves Carvalho Barbosa (Seção de Gestão de Sistemas de Gravação-SESIG) via email, January 30, 2019. 

6A similarly well-executed project by a Latin American government (although not related specifically to politics) is Mexico’s Fonoteca Nacional, which has created an archive of audio materials judged to be part of the nation’s cultural heritage. Working with state and local governments, the Fonoteca Nacional has established 108 fonotecas virtuales from which its archive may be accessed in cultural and educational institutions across Mexico, with at least one in every state except two ( Alejandro L. Madrid (2006) has recently problematized the Fonoteca’s exaltation of itself as a project that democratizes access to sound. 

7For example, for early feminist historians, audio recordings of interviews offered unmediated access to voices that patriarchal societies had long ignored (Gluck 2014, 35–37). 

8For a recent study that uses voice and body language to complicate the analysis of an oral history interview and produce a more nuanced transcription, see Cheng 2018. 

9For an analysis of how the questions asked by oral history have changed with the advent of audio, video, and online/digital technologies, see Larson 2016. 

10For a discussion of the hermeneutics of oral history, particularly the embodied nature of the interview, see Friedman 2014. 

11Diário da Câmara dos Deputados, Diário do Congresso Nacional, and Diário da Assembléia Nacional Constituinte are official publications of the Câmara dos Deputados that contain transcripts of specific types of congressional sessions. See 

12The statements cited are from Carlos Santana (PT-Rio de Janeiro), Francisco Silva (PDC-Rio de Janeiro), and Maria Laura (PT-Distrito Federal). 

13In 2016, the deputies’ dedications of their votes to their grandchildren, their constituents, God, and a variety of pet causes were widely ridiculed in the media. Nevertheless, the significant continuities with the declarations made during the 1992 impeachment vote indicate that Brazilian politicians have their own notions about honor, duty, and legitimacy that appear strange or even silly to ordinary Brazilians, but which are transmitted from one generation of politicians to the next by elite families and the institutional culture of Congress itself (Sims 2016; Aquino 2016). 

14Vilma Pereira, interview with Bryan Pitts, Brasília, April 10, 2015. 

15Due to server space limitations, the Câmara only offers the last one to two years of video on its site, which means that this speech is no longer available. The video was, however, uploaded to YouTube: “Homenagem ao Dia da Empregada Doméstica. Dep. Benedita da Silva (PT-RJ),” YouTube video, 11:28, posted April 29, 2014, 

16The following two examples from student papers are reproduced with the permission of the authors. 


We would like to thank the LARR editors and readers for their feedback on this research report. In addition, Daniel James and the graduate students in HIST-H 699 (Colloquium in Oral History) at Indiana University Bloomington offered comments on an earlier version of this manuscript.

Author Information

Bryan Pitts is Assistant Director of the Latin American Institute at the University of California, Los Angeles. A historian of Brazil by training, he is completing a book manuscript on the role that both military tutelage and popular mobilization had in reshaping the culture of the civilian political class during the country’s 1964–1985 military dictatorship. He has also conducted research on race and LGBT media in Brazil and gay Brazilian tourists’ experiences of race, nationality, and sexuality while traveling abroad. He writes extensively on contemporary Brazilian politics for media outlets in both the United States and Brazil.

Yahn Wagner Pinto is an Assistant Professor of Composition at the Universidade Federal do Rio de Janeiro, a composer, and a doctoral candidate in Composition in the Department of Music at Duke University. His work has been played and recorded by groups like the Villa-Lobos Quintet Ensemble, one of Brazil’s most respected chamber music groups. He has twice been awarded the Circulation of Concern Music Prize from the Fundação Nacional de Artes (Funarte), which allowed him to write seven chamber pieces and play them in sixteen Brazilian states. His research is dedicated to the phenomenon of creativity in music composition.

Madeleine Roberts traces her interest in Brazil to a “Global Brazil” course she took with Bryan Pitts at Duke University. She proceeded to learn Portuguese and spent five months traveling across Brazil. She wrote her senior thesis in History on the history of and social discourse around domestic workers in Brazil. As part of her research, she interviewed empregadas domésticas in Fortaleza, analyzed thousands of classifieds on, and researched the legislative history of Brazil’s 2013 constitutional amendment recognizing the rights of domestic workers. Madeleine now works for a real estate developer in Washington, DC, but she profoundly misses academia.


  1. Aquino, Wilson. 2016. “Discursos de deputados na votação causaram constrangimento em eleitores.” O Dia. April 19. 

  2. Cheng, Nien Yuan. 2018. “Flesh and Blood Archives: Embodying the Oral History Transcript.” Oral History Review 45 (1): 127–142. DOI: 

  3. Cleveland, Daniel. 2017. “Sound Analysis: Pará.” Assignment for PORT 4040 (Brazilian Politics and Political Culture), University of Georgia. 

  4. Friedman, Jeff. 2014. “Oral History, Hermeneutics, and Embodiment.” Oral History Review 41 (2): 290–300. DOI: 

  5. Gluck, Sherna Berger. 2014. “Why Do We Call It Oral History? Refocusing on Orality/Aurality in the Digital Age.” In Oral History and Digital Humanities: Voice, Access, and Engagement, edited by Douglas A. Boyd and Mary A. Larson, 35–52. New York: Palgrave Macmillan. DOI: 

  6. Larson, Mary A. 2016. “‘The Medium Is the Message’: Oral History, Media, and Mediation.” Oral History Review 43 (2): 318–337. DOI: 

  7. Lemos, Cláudia Regina Fonseca, Antônio Teixeira de Barros, and Cristiane Brum Bernardes. 2016. “Public Communication in the Brazilian Congress: The News Agency and TV Station of the Chamber of Deputies.” Latin American Research Review 51 (4): 202–224. DOI: 

  8. Madrid, Alejandro. 2016. “Landscapes and Gimmicks from the ‘Sounded City’: Listening for the Nation at the Sound Archive.” Sound Studies 2: 119–136. DOI: 

  9. Pitts, Bryan. 2011. “‘O funeral da democracia’: O caso Moreira Alves, a cultura política das elites e o estabelecimento de uma ditadura no Brasil.” In Paulistânia eleitoral: Ensaios, imagens, dados, edited by José D’Amico Bauab, 285–300. São Paulo: Tribunal Regional Eleitoral, Imprensa Oficial. 

  10. Pitts, Bryan. 2013. “The Inadvertent Opposition: The São Paulo Political Class and the Demise of Brazil’s Military Regime, 1968–1985.” PhD dissertation, Duke University. 

  11. Roberts, Madeleine. 2017. “Domestic Work in 21st Century Brazil: From ‘a Foot in the Kitchen’ to a ‘Foot in the Door.’” BA honor’s thesis, Duke University. 

  12. Sims, Shannon. 2016. “The Most Outrageous Moments of Brazil’s Impeachment Hearing.” Forbes, April 18. 

  13. Wagner, Yahn. 2017. Magina: In Memoriam, Mario Juruna (for Septet and Tape). 

  14. Zhu, Lilian. 2017. “Sound Analysis: São Paulo.” Assignment for PORT 4040 (Brazilian Politics and Political Culture), University of Georgia.