Rodrigues, Aldair. "African Voices in Manumission Records: Freedom Certificate Data Extracted
from livros de notas and livros de alforrias (Mariana, Brazil, 1711-1807)." Journal of Slavery and
Data Preservation 4, no. 4 (2023): 45-55. https://doi.org/10.25971/ywwf-1c82.
African Voices in Manumission
Records: Freedom Certi cate
Data Extracted from livros de
notas and livros de alforrias
(Mariana, Brazil, 1711-1807)
Peer-Reviewed Dataset Article
Article Authors
Aldair Rodrigues, UNICAMP
Dataset Creators
Aldair Rodrigues, UNICAMP
Contributors
Caroline Cunha, UNICAMP
Franceline Galdino, UNICAMP
Juliana Soares, Universidade Federal de Ouro Preto
Leonardo Leoz, UNICAMP
Lucas Samuel Quadros, Universidade Federal de Ouro Preto
Lyandra Amaral, UNICAMP
Natã Freitas, UNICAMP
JSDP (ISSN 2691-297X) 4: 4 (2023)
45
Journal of Slavery and Data Preservation - Freedom Certificate Data, Mariana, Brazil Peer-Reviewed Dataset Article
Pedro Gericó, UNICAMP
Renata Diório, University of São Paulo
Victor Sampaio, UNICAMP
_________________________________________________________________________________________
Description
This dataset contains information about 3,260 freed people who lived in Mariana (Captaincy of
Minas Gerais), in the Brazilian mining region, during the eighteenth century. The data were
extracted from two archival series belonging to Arquivo Histórico da Casa Setecentista de
Mariana. Most of it comes from 92 livros de notas (notary book of notes); the first códice dates
back to 1711, when the town was founded, and the last one documents information from 1770.1
These sources are complemented by two livros de alforrias (freedom certificate books)
registering manumissions.
The register livros de notas were used by the Crown official notary to record a variety of
contracts and deeds, ranging from trading partnerships to bills of sale for land, enslaved people,
gold mines, etc. Our objective is to facilitate access to data about manumissions that are not
registered in dedicated codices and therefore are scattered across multiple types of documents
and not cataloged . Headings that indicate a manumission in livros de notas include: escritura de
alforria e liberdade (deed of manumission and freedom), escritura de liberdade (deed of
freedom), carta de liberdade (letter of freedom), papel de liberdade (freedom paper), and escrito
de liberdade (freedom writing), among others. Douglas Lima explains that each type of freedom
certificate and their bureaucratic details had specific implications in the lives of Black people
shaped by power relations and social control strategies.2 Despite these variations, the core of
their legal impact tends to align. The documents provide the similar basic information that
became variables in our spreadsheet, such as the name of the manumitted person, his or her
origin, enslaver who granted the manumission, type of concession, and payment and
justification for the act of manumission.
Unfortunately, we were not able to extract data from livros de notas beyond 1770 (although they
continued to be used until the end of the eighteenth century) due to lack of resources. In order
to still somewhat allow researchers to access information about the late 1700s, we compiled
data from two codices dedicated only to record manumission – livro de alforria 2 (1740-1794)
and livro de alforria 3 (1792-1808). Those are the only codices explicitly about manumission for
the timeframe 1740-1808. Pairing data from two different kinds of notary documentation, which
This period coincides with the timeframe of a larger project dedicated to examining the knowledge
production about the origins of Africans who lived in Minas Gerais.
2
Douglas Lima, Libertos, patronos e tabeliães : a escrita da escravidão e da liberdade em alforrias
notariais (Belo Horizonte, MG: Caravana, 2020), chap. 2.
1
JSDP (ISSN 2691-297X) 4: 4 (2023)
46
Journal of Slavery and Data Preservation - Freedom Certificate Data, Mariana, Brazil Peer-Reviewed Dataset Article
chronology partially coincide (1740-1770), allows researchers to evaluate how representative
livros de alforrias are. It would be possible, for instance, to check if a freedom document
Fig. 1: carta de alforria (freedom certificate). Source: Livro de Notas, n. 9 (1717), Arquivo Histórico da
Casa Setecentista de Mariana.
registered in livro de notas is also registered in livro de alforrias. The logic underlying the
registrations is not clear, since the livros de alforrias tend not to repeat the freedom certificates
found in livros de notas. In summary, we do not recommend statistical use of this dataset
without taking into consideration these limitations in the archival sources. It could risk
producing distortion in the perception of its general trends.
Our ultimate goal is to disseminate qualitative information containing biographical data on freed
people centered on their name. The dataset can shed new light on the demographics and
JSDP (ISSN 2691-297X) 4: 4 (2023)
47
Journal of Slavery and Data Preservation - Freedom Certificate Data, Mariana, Brazil Peer-Reviewed Dataset Article
agency of Africans and their descendants who became freed people in colonial Brazil,
documenting their name, origins, and important aspects about the negotiation process that
culminated in their new legal status. It can also reveal how the blurred lines between
enslavement and freedom permeated their experience within the African diaspora.
Fig. 2: Minas Gerais in the eighteenth century. Source: Claudia Fonseca, “Urbs e civitas: a formação dos
espaços e territórios urbanos nas minas setecentistas,” Anais Do Museu Paulista: História e Cultura
Material 20, no. 1 (2012): 84. Reprinted with permission of the author.
Dates of Data Collection
2014-2021
Dataset Languages
Portuguese, English
JSDP (ISSN 2691-297X) 4: 4 (2023)
48
Journal of Slavery and Data Preservation - Freedom Certificate Data, Mariana, Brazil Peer-Reviewed Dataset Article
Geographic Coverage
Minas Gerais, Mariana, Pernambuco, Bahia, Espírito Santo, Rio de Janeiro, Portugal, Madeira,
Itália, Argel, Mandinga, Guinea, Cabo Verde, Costa da Mina, Chamba, Bight of Benin, Dahomey,
Alada, Savalou, Nigeria, Bight of Biafra (Calabar, Carabali), Congo, Loango, Angola, Luanda,
Massangano, Quiçama, Benguela
Temporal Coverage
1711-1807
Document Types
Emancipation certificates
Notary records
Sources
Livros de Notas. Arquivo Histórico da Casa Setecentista de Mariana. Mariana, Brazil.
Cartas de Alforrias. Arquivo Histórico da Casa Setecentista de Mariana. Mariana, Brazil.
Methodology
The project’s first step was the digitization of 92 livros de notas and two codices titled cartas de
alforrias (manumission letters) in the Casa Setecentista de Mariana archive. We next trained a
team of students to collect data in a spreadsheet, which was subsequently assessed and
revised by the dataset coordinator.
The main challenges faced by the team at the beginning was, first, acquiring the paleography
skills to read eighteenth-century Portuguese manuscripts. Second, we had to deal with material
conditions of the documents. When part of the document was damaged, illegible, or missing, we
compiled the remaining information into the spreadsheet, indicating the lacking variables as
“corroído” (eroded) between square brackets; sometimes most of it was added, making it
impossible to consider the record in our dataset. Some códices have a hole in the middle of the
pages, probably bored by dripping water. Fortunately, though, the vast majority of the series
livros de notas are in good condition in terms of material preservation.
Each row in the spreadsheet corresponds to a person and displays their information organized
in columns (one for each variable, as explained below). The same letter of manumission could
emancipate more than one person, such as a mother and her children or spouses. In such
cases, we created a new row for each person, repeating the information pertaining to all of
them in each corresponding record.
JSDP (ISSN 2691-297X) 4: 4 (2023)
49
Journal of Slavery and Data Preservation - Freedom Certificate Data, Mariana, Brazil Peer-Reviewed Dataset Article
The rationale behind the spreadsheet design was shaped by two main objectives: a) first, to
enable researchers to recover names and fragmented biographical information regarding the
agency and voices of the enslaved person who was being freed; b) second, make it possible to
trace someone’s trajectory through further research in colonial archives, either for academic
purposes or for broader public usage, such as genealogical inquiry. That is why special attention
has also been given to designing the spreadsheet to facilitate further disambiguation strategies
using origin and enslaver name. This project might contribute to Brazilian citizens of African
descent finding answers to where they came from.
In most manumission records, we find only the first name of the person being manumitted.
Later they would adopt a surname, which tended to be acquired from their former owner’s family
name, following the Roman patronage traditions observed in the Portuguese settings. Their
Christian given name would be accompanied by origin information (nação, nation), such as
Maria Mina or Maria Angola, or skin color associated with social status (qualidade), like João
pardo or Ana mulata, etc. Searches by name can be therefore ambiguous. A recommended
strategy to overcome this would be referencing first name along with origin and/or with the
enslaver’s name who granted the manumission.
It is worth noting that the ambiguous results from name-based searches come about when
enslaved people are digitally represented in spreadsheet cells. In the colonial context of the
eighteenth century, however, each person’s place in the social relations of the parishes in which
they lived and their physical characteristics would easily distinguish them; enslavers, overseers,
neighbors, companions, and friends would easily differentiate one person from another. Few
elements of these social relations were recorded in the documents. One must be aware,
therefore, that a dataset enables the retrieval of only fragments that characterized a person in
the wider social fabric in which they lived. Though fragments, they are absolutely crucial to the
historical understanding of those realities and enslaved peoples’ lives. We believe that the
digital treatment of loose threads and traces left by those experiences in the colonial archive
enhances the possibilities of interrogating that reality.
Origin descriptor, denoting nation and qualidade (social status), is the most complicated
variable in the dataset. First, there were several ways of classifying African origins and their
spelling. Second, a notary could use crioulo as nation to refer to Brazilian-born people or to
people who came from other parts of the Portuguese empire, such as crioulo de Angola, crioulo
da Madeira, or mulato de Portugal. The clerk would write, in these cases, nação crioulo or nação
crioula. Third, notaries often substituted nation with social status (“qualidade”) in documents
pertaining to mulatos and pardos, equating origin with skin color categories. Fourth, the variable
also included indigenous background, especially in the first decades of the eighteenth century.
Indigenous people appear as gentio da terra (gentile of the land, meaning native Brazilian
indigenous) or gentios do cabelo corredio (gentiles with straight hair), as opposed to gentio da
guiné (gentile from Guinea, meaning African-born). To compile the origin variable, we considered
the contemporary logic that would lead a clerk to insert these various types of information with
someone’s first name. That is why they occupy the same column of the spreadsheet. To further
JSDP (ISSN 2691-297X) 4: 4 (2023)
50
Journal of Slavery and Data Preservation - Freedom Certificate Data, Mariana, Brazil Peer-Reviewed Dataset Article
complicate matters, a clerk could also use two different terms to talk about someone’s ethnicity
or race, applying, for example, mulato in one part of the document and switching to pardo in
another mention in the same manumission record; in such cases, we kept both.
There were multiple layers of meanings related to nation (nação) when used to talk about
African origin. Sometimes it designates toponyms of the macro-areas of organized slave trading
on the African coast, ports of shipment, polities, identities connected to political affiliations;
other times it designates ethnonyms referring to more specific ethnic identities. Depending on
the moment of someone’s biography and the context in which a person was described, nation
could vary between a subgroup and broader generic descriptor.
The spelling of the African origin nomenclature also varies, even when it refers to the same
person. We chose to keep the original form in a first field (Origin/Nation) exactly as it appears in
the documents, and created a second column to standardize nations and enable their
serialization and possible encoding. The methodological reason for maintaining the
nomenclature variations are their historical relevance, especially to ethnolinguistics. They may
encapsulate multiple dynamics and interactions between slave trade agents and the narratives
of Africans themselves about their origins (kingdoms, ethnolinguistic designation, lineage, etc.).
They could designate their conception of belonging based on ancestry and territoriality, since
several terms clearly reflect African pronunciations of the lexicon that described their origins. As
new terms being incorporated into the colonial world, scribes often tended to write them down
as they listened but adapting terms to Portuguese characters. More specific terms of African
origin appeared concurrently or overlapped with more generic descriptors attributed by
merchants, such as "Mina" meaning Costa da Mina in West African Coast.
In terms of demographic features underlying the documents’ production, we are learning about
a high density of people from the same African region. During the Brazilian gold rush of the first
half of the eighteenth century, the urban setting of the Minas Gerais mining region became
home to one of the largest concentrations of people from the Bight of Benin in the Americas.
Most of them were men who had been enslaved during the wars related to the expansion of
Dahomey kingdom, which advanced from the Abomey plateau to the Atlantic, encompassing the
ancient kingdoms of Alada in 1724 and Whydah in 1727.3 Many individuals could be and were
described with the same combination of first name and the origin descriptor.
One telling example from these documents is the origin of the "Ladano" nation. It is probably the
Portuguese transliteration of Alladahonu, which in the Gbe area of West Africa meant people
from Alada.4 It connotes the political affiliation of the subjects of the old kingdom of Alada, in
the southeast of current Benin. The spelling of this term oscillated between "Ladano" and
Robin Law, The Slave Coast of West Africa 1550-1750: the Impact of the Atlantic Slave Trade on an African
Society (Oxford: Clarendon, 1991).
4
Luis Nicolau Parés, O Rei, O Pai e a Morte: A Religião Vodum na Antiga Costa dos Escravos na África
Ocidental (São Paulo: Companhia das Letras, 2016), 48; Anatole Coissy, "L'arrivée des 'Alladahonou' à
Houawe,'" Études dahoméennes 13 (1955): 33-34.
3
JSDP (ISSN 2691-297X) 4: 4 (2023)
51
Journal of Slavery and Data Preservation - Freedom Certificate Data, Mariana, Brazil Peer-Reviewed Dataset Article
"Ladana," a gendered adaptation to the Portuguese language. All these variations were recorded
in the spreadsheet to digitally maintain the layers of this historical process. In the case of
Ladano Ladana, the fluctuations were updated to "Alada," which was the most common term
used in the European world to designate that political unit.5
The variable origin can offer important insights on slave trade dynamics and routes, within the
African continent and routes that dislocated crioulos, pardos, and mulatos from Pernambuco,
Bahia, and Rio to fulfill the gold rush’s demand for labor.
Upon these methodological reflections, we keep these original forms in one field ("origin/
nation") and in a second field control them as follows:
a) standardize variations of the nomenclature for the term according to current African
forms, when it can be established;
b) update it to contemporary Lusophone spelling, when these words are still used;
c) or, if the two options above are not possible, to standardize oscillations for the most
recurring term in the documentation.
ORIGINAL SPELLING VARIATIONS
Ladá, Lada, Ladano, Ladanu, Ladana
Anagô, Anagó, Anago, Anagonu, Naguno,
Nagono
Fom, Fono, Fona, Fon
Sabarú, Sabará, Sabalu
Courá, Courano, Courana
Xambá, Chambá
Cravari, Crabari, Craballi, Caravali, Carabali
Engola, Engolla, Angolla, Angola
Banguela, Banguella, Benguella, Benguela
STANDARDIZED VERSION
Alada
Nagô
Fon
Savalou
Courá
Chambá
Carabali
Angola
Banguela
The variable “payment” is important because it reveals nuances about the agency of enslaved
people who acquired freedom. The ways of obtaining alforria can be grouped into two main
strategies: free or paid. In the first case, the enslaved did not have to offer pecuniary payment
for his or her freedom. Data about this process can be observed in the field “Reason for the
manumission concession.” Alternatively, we found multiple ways to achieve freedom through
purchase, either by enslaved persons themselves or by third parties, such as partners, mothers,
godparents, etc.
Robin Law, The Kingdom of Allada (Leiden: Research School CNWS, School of Asian, African, and
Amerindian Studies, 1997). These aspects related to Minas Gerais are further contextualized and
analyzed in Aldair Rodrigues, “‘Com Duas Gejas Em Cada Uma Das fontes’: escarificações e o processo
de tradução visual da diáspora jeje em Minas Gerais durante o século XVIII,” Afro-Ásia 63 (June 2021),
https://doi.org/10.9771/aa.v0i63.38662.
5
JSDP (ISSN 2691-297X) 4: 4 (2023)
52
Journal of Slavery and Data Preservation - Freedom Certificate Data, Mariana, Brazil Peer-Reviewed Dataset Article
Manumission could be purchased using cash, gold, or another enslaved person. In the first two
means, it could be paid in installments (quarterly, quartação) or all the amount at once. The
Portuguese terms vary from cortado/cortada to quartado/quartada, all deriving from the verb
quartar (to divide in quarters, though in reality the installments were not exactly a quarter).
Eduardo França Paiva and Andrea Lisly Gonçalves emphasize that quartação was one of the
most common ways of self-purchasing manumission in the urban settings of Minas Gerais.6 We
either transcribed or summarized the details regarding this process in this field in order to reveal
the agency of enslaved people in their quest for freedom.
In the context of Brazilian urban slavery, an enslaved person could hold ownership over another
enslaved person, especially when they lived as escravos de ganho (slaves for hire), paid for
various services they provided for different people. João Reis, who thoroughly examined this
complex phenomenon in the Bahian context, defined this kind of arrangement as alforria de
substituição (substitution manumission).7
In this dataset, a small percentage of manumissions were paid with another enslaved person:
35 cases. We transcribed the information about the person being used as payment in this field.
It is not always clear if the person being manumitted acquired the enslaved only for the purpose
of substituting lost human property. In some cases, the enslaver clearly indicates that the
person being manumitted held ownership of the person while being enslaved himself for a long
time, specifying that they either used that person as payment or to make clear they were
allowed to keep that enslaved person after themselves being freed. Other times, they simply
mentioned that the manumitted possessed an enslaved person, regardless of them being used
as payment. Understanding this complicated phenomenon requires further research in
additional documents, especially those that provide more qualitative information, such as
testamentos (wills) or court proceedings.
We should bear in mind that the information collected from the colonial archive was generated
amid social relations that defined who could speak in the notary’s office.8 The catalogs of the
colonial archives replicate the centering of the agency of slave owners and the ways they
created and organized information. A simple indication of that is provided by the archival series
registros de testamentos (testament wills register), where we can read documents in which the
voices of freed Africans prevail, despite the colonial filter of the official scribes that permeated
them. We found 99 testament wills belonging to freed people. The same person had a
Eduardo Paiva, “Coartações e alforrias nas Minas Gerais do século XVIII: as possibilidades de libertação
escrava no principal centro colonial,” Revista de História (USP) 133 (1995): 49-57; Andréa Lisly
Gonçalves, As margens da liberdade: estudo sobre a prática de alforrias em Minas colonial e provincial
(Belo Horizonte: Fino Traço, 2011). This practice was also widespread in the Spanish empire, as observed
in Cuba. See: Alejandro de la Fuente and Ariela Gross, Becoming Free, Becoming Black : Race, Freedom,
and Law in Cuba, Virginia, and Louisiana (Cambridge: Cambridge University Press, 2020), 105-114.
7
João José Reis, “Por sua liberdade me oferece uma escrava”: alforrias por substituição na Bahia,
1800-1850,” Afro-Ásia 63 (June 2021), https://doi.org/10.9771/aa.v0i63.43392.
8
This discussion is inspired in part by Ann Laura Stoler, “Colonial Archives and the Arts of Governance,”
Archival Science 20, no. 1 (2022): 87-109.
6
JSDP (ISSN 2691-297X) 4: 4 (2023)
53
Journal of Slavery and Data Preservation - Freedom Certificate Data, Mariana, Brazil Peer-Reviewed Dataset Article
manumission certificate registered in livros de notas or livros de alforrias in only eighteen cases.
9
That preliminarily indicates that the majority of their manumission certificates were not
recorded in notary registers (livros de notas or livros de alforrias).
Further research on documentation produced by their enslavers could reveal how it was
processed in terms of legal effects. We found evidence in this documentation about three freed
persons who contested information about the content of their manumission certificate (carta de
alforria). One of them is Maria de Meira, from Costa da Mina (Bight of Benin, Slave Coast). In
1751, she insisted that it was not true her freedom had been given free of charge by her former
owner Manoel Marques, as stated in the notary archive. She claimed instead that her freedom
was purchased: "I paid him two pounds of gold, as shown in the receipts that I have kept along
with my letter of manumission. Though it says the letter is given free of charge, I owe nothing to
my mentioned master."10 Unfortunately, we have not found her carta de alforria in livros de notas,
but this example demonstrates the importance of cross-examining information on the same
person from different angles in order to shed light on power relations underlying the production
of legal information in slave society archives.
Although thorough, this dataset does not encompass all the manumissions granted in Mariana
during the period covered by this project. First, it does not necessarily include freedom given in
testament wills or in baptismal records, which were legally valid despite not being recorded and
kept in the livros de notas or de alforrias. Only cross-examining data collected in wills or parish
records with livros de notas will reveal the percentage of coincidence of the records belonging to
the same person, which is beyond the scope of this article and dataset. Finally, one should be
conscious that the vast majority of Africans brought to Brazil died enslaved.
Dataset users should read critically the information it contains considering it was produced in a
context of a slave society that commodified and racialized Africans and people of African
descent through derogatory terms equating skin color, social and legal status. That can be seen,
for instance, in words like mulatinho, mulatinha, crioulinho, crioulinha, cabra, cabrinha, etc. Any
use of this kind of terminology today should be accompanied with a contextualizing
explanation. In this data article we have worked to provide information to that end.
Notes on transcription and extraction:
● Blank: Blank cells are used when the expected information is not available in the primary
source.
Arquivo Histórico da Casa Setecentista de Mariana, Registros de Testamentos, códices n. 50, 53, 57, 59,
60, 62, 63, 69, 65, 67, 71, 72, 73.
10
Original in Portuguese: “me quartou em duas Libras de ouro que lhe paguei como consta dos recibos
que tenho do dito junto a minha carta de alforra que suposto a dita carta está passada gratuitamente não
devo nada ao dito meu senhor”. Arquivo Histórico da Casa Setecentista de Mariana, Registros de
Testamentos, 1751, Livro 71, fl. 121v.
9
JSDP (ISSN 2691-297X) 4: 4 (2023)
54
Journal of Slavery and Data Preservation - Freedom Certificate Data, Mariana, Brazil Peer-Reviewed Dataset Article
●
●
●
Eroded [corroído]: This indicates when information is in the document but we cannot
read it due to material deterioration of the codice.
Illegible [ilegível]: Some words could not be read due to several factors, such as faded
pen ink, undecipherable handwriting, bleedthrough, etc. In these cases, we inserted the
word [ilegível] (illegible) between square brackets.
Use of square brackets: When we are not sure about the manuscript transcriptions, we
put the word or part of phrases between square brackets [ ].
Date of Publication
September 2023
Data Links
Dataset Repository: Harvard Dataverse, https://doi.org/10.7910/DVN/L9TWHK
Linked Data Representation: Enslaved.org
Acknowledgments
Prêmio Capes AUXPE 0382/2016, 23038.009186/2013-63
CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) - Bolsa Produtividade
Processo 309752/2021-3
FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo) - 2014/23508-2
UNICAMP/ Serviço de Assistência Social/ Bolsa Auxílio Social
UNICAMP/ FAEPEX 3201/16 (Auxílio Início de Carreira Docente)
CECULT (Centro de Estudos sobre História Social da Cultura)
Cite this Article
Rodrigues, Aldair. "African Voices in Manumission Records: Freedom Certificate Data Extracted
from livros de notas and livros de alforrias (Mariana, Brazil, 1711-1807)." Journal of Slavery and
Data Preservation 4, no. 4 (2023): 45-55. https://doi.org/10.25971/ywwf-1c82.
Copyright
© 2023 Aldair Rodrigues. This article is distributed under the terms of the Creative Commons
Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0), which permits
non-commercial use, distribution, and reproduction provided the original creator and source are
credited and transformations are released on the same license. See
https://creativecommons.org/licenses/by-nc-sa/4.0/.
JSDP (ISSN 2691-297X) 4: 4 (2023)
55