Freedom Narratives respects Best Practices as developed in the course of generating the website and database. In addition to the identification of the specific contributions of the Research Team, these Best Practices include a description of the Methodology employed in data management, the Controlled Vocabulary that has been devised to enable the management of clean data for ingestion into the Project, including Definitions of terms and vocabulary, and the suggested format for Citation of the database, website and their constituent parts. Best Practices are described in Paul E. Lovejoy, Érika Melek Delgado and Kartikay Chadha, “Freedom Narratives of West Africans from the Era of Slavery,” which is summarized in the following sections.
The Freedom Narratives website is an open source relational database comprised of original documentation in PDF format with metadata organized into data fields. The database facilitates access to all documentation through this public website. The intention is to enable the analysis of the important historical tragedy and crime against humanity that affected the history of West Africa and was responsible for the demographic, cultural and social transformation of the Americas and elsewhere. The focus on biographical profiles of people in West Africa during the era of the trans-Atlantic slave trade is an innovative approach to social history. The Project is based on an online digital repository of autobiographical testimonies and biographical data of Atlantic Africans that allow the analysis of patterns in the slave trade, specifically in terms of where individuals came from, why they were enslaved, and what happened to them. Because Freedom Narratives focuses on those born in Africa, in most cases people had been born free and therefore are to be distinguished from those who were born into slavery in the Americas or elsewhere. The important distinction between those born free from those born into conditions of slavery is highlighted by reference to their testimonies as “freedom narratives” rather than as “slave narratives,” the term that is commonly used to categorize the personal accounts that have been recorded in the Americas. In the case of the surviving accounts of those born in West Africa, at least, not only were most individuals born free but were subsequently able to regain their freedom. The experience of slavery was therefore only an aspect, even if an important aspect, of their personal experiences. By contrast, those born into slavery in the Americas and elsewhere were socialized into slave societies from birth, and even if individuals were able to acquire their freedom, their exposure to slavery was different. The accounts that are being labelled “freedom narratives” often recount the hope, sometimes realized, of being reunited with kin or members of natal societies that was more difficult and often impossible in the trans-Atlantic and trans-Saharan diasporas. Individuals who were never enslaved are also included because of the importance of identifying various kinds of relationships, not only those based on kinship and social interactions but also relationships of dependency and subjugation. Freedom Narratives enables an examination of biographical testimonies as the fundamental units of analysis in historical reconstruction, whether the primary texts arise from first person memory or survive via amanuensis. Whenever possible, original testimonies are supplemented with biographical details culled from legal, ecclesiastical, and other types of records. Through its website, the Freedom Narratives project assembles, collates and displays biographical information on individuals who were primarily born in West Africa or returned to West Africa during the centuries of the slave trade as a means of understanding the episodes and trauma of the experience of slavery, whether the enslaved crossed the Atlantic to the Americas, remained in West Africa or were taken across the Sahara.
Freedom Narratives uses the Decoding Origins Web Portal (http://decodingorigins.org), which consists of a backend collaborative database built using Structured Query Language (MySQL) and a user friendly interactive platform employing novel algorithms to collect, analyze and cross reference data. The database is designed to handle multiple types of documents from a variety of sources that often constitute different categories. This web portal employs user-sensitive inclusive design and iterative learning methodologies to improve the data curation process not only for Freedom Narratives but adaptable for other projects and compatible with the Enslaved hub. Because the Decoding Origins web portal provides a project-customized, collaborative environment to collect, organize, meta-tag, visualize and cross-reference data and connect with external collections, the portal can also connect with public websites to enable real time publications of biographies and documents. The Decoding Origins Web Portal allows historians to work in a team-based web-environment that reinforces the rigorous analog best-practices developed by the Freedom Narratives research team. Instead of alteration between documents, the Decoding Origins Web Portal allows researchers to link PDF documentation and corresponding type work with Controlled Vocabularies, simultaneously displayed on one screen. The Source tab of the Decoding Origins Web Portal records the name of the digital files that are being processed. The basic structure of data management is based on four interrelated methods of identification, first establishing the “Person” as a unique individual, then examining the various known “Events” in that person’s life, and the “Places” where such events occurred, the “Sources” on which the data are based. These four dimensions of management are organized digitally as an open source relational database.
The “Person ID” entry is a unique identifier for each case file entered in the Freedom Narratives database. Researchers are assigned a specific range of identifiers, which are called “FN Numbers.” These identifiers follow the same format: FN, followed by six digits (e.g., FN000001, FN001200, …), which allow differentiating one individual from another who might have the same name. Only a project director can create a new person file on the Decoding Origins Web Portal and then a “FN number” is automatically assigned to each case file for a new person. Assigned cases are listed under “My Task” section of the web portal for purposes of work flow.
The “Name” entry records the name of the individual as it is reported in the primary source document. Whenever possible, the African name of the individual is used. If this is not possible, the most common name is used, including every part of the name as it is written in the source: given/first name, surname/last name (e.g., André do Amaral). If the primary sources do not mention the individual’s name, the person is described as Unknown. Sometimes an individual is referred to by more than one name. Such “Alternative Names” can be name variations, nicknames, aliases, married/maiden names, assumed names, variant spellings of African and Christian names, alternative spellings with abbreviations or initials, and a number of other things. A semicolon ( ; ) is used to differentiate multiple entries in various fields of the Web Portal or a multiple select HTML form type field is provided.
The identification of individuals usually includes reference to an ethnonym or language, which for purposes of data entry and analysis is distinguished according to the term(s) that are recorded in the primary documentation and through association with modern terminology as appropriate. The primary sources usually record the “ethnonyms” of an individual as ethnicity, tribe or nation. The assumption that is made for Freedom Narratives is that such references are most likely to the language(s) an individual speaks. An “Ethnonym / Language in Original” entry is recorded exactly as in the primary documents (e.g., Poullar, Fula, etc.). These terms are then sifted through the Controlled Vocabulary, which is an ever-growing list containing terminology derived from Sources and includes variations of ethnonyms in primary sources, modern scholarship and common usage. Ethnonyms are linked to likely region of origin (e.g., Fulbe | Poullar; Pualard; | Upper Guinea Coast or Senegambia), that is the broad area from where an individual originally came. If the primary document does not specify where the individual was born, the “Ethnonym / Language” Controlled Vocabulary can help approximate the location. As an example, if FN00XXXX is identified as Allada, the “Region of Origin” is the Bight of Benin region. Region of origin is determined with reference to the regionalization outlined in Henry B. Lovejoy, Paul E. Lovejoy, Walter Hawthorne, Edward A. Alpers, Mariana Candido, Matthew S. Hopper, “Redefining African Regions for Linking Open-Source Data,” History in Africa 46 (2019), 5-36.
The region of destination records the last location where an individual is noted in the sources, and ultimately, the location where he/she died, if known. For purposes of Freedom Narratives, regions of destination include the following designations: Caribbean, North America, western Europe, Hispanic mainland, Brazil. As an example: the “Region of Destination” for an individual found in the Marronnage documents would often be “Caribbean.” If it is later discovered that a person travelled to, or died in, North America, this new information would replace “Caribbean.”
The biological sex of individuals as listed in the source document(s), often based on observations, is included as female, male, or unknown. All information about the physical appearance of an individual in primary documents are grouped according to characteristics under specific terms (e.g., wound; branding; height). Individual family relationships are considered part of an individual’s identity, and hence parents, children, spouses, and other close family are identified. It is recognized that individuals virtually always understood more than one language, including languages of interrogation or interviews as well as those associated with ethnonyms, and in some cases literacy is indicated, such as when the primary documents state that the individual could write, or if it is possible to conclude that he/she could (e.g., if a document was written by this individual in a certain language).
A basic summary of what is known about an individual is posted on the website for easy reference. The template for summaries includes basic information, including the name of the individual, his/her sex, identification of ethnonym/language, an estimate of the year or range of years in which the person was born, where the individual was born specifically as to place and geographic region. When, where and how an individual was enslaved is noted along with when and where the individual was taken and what happened later, such as emancipation, sale, escape, marriage and any unique details. Finally, the summary includes the year in which the information was recorded, who obtained the information, where this was done, and the final location of the individual if known.
If information is also to be found in a contributing project, such as Slave Voyages, Liberated Africans, Harvard Biographies or Le Marronnage dans le monde atlantique, the ID number that identifies the document/information in that project’s database is noted, which for the Slave Voyages database means the “Voyage ID” number; for Le Marronnage database the ID number for each slave advertisement (at the end of each document’s URL), and for Liberated Africans, the registration number in the original documentation. Similarly, every person contributing to the project receives proper recognition for his/her work, which includes those who provided the data and those who entered the data in the database.
The events of a person’s life are recorded in a similar fashion, following Best Practices and relying on Controlled Vocabularies that have been determined for the project. An “Event Type” is a category or class that captures an event’s overarching impact or purpose, as noted in the primary sources. Some will be obvious since there is an “Information Recorded” event when the document is produced. Others require more reflection; for example, if a runaway African recorded in a Les Affiches Américaines newspaper is referred to by a French name, it can be assumed that there was a “Naming Ceremony” event, including baptism, at some point in the individual’s personal history which had to have taken place before the “Information Recorded” event, and even before the “Resistance” event (when the person ran away from the master). Information is entered chronologically based on the Controlled Vocabulary for “Event Types.”
The social position or status of the individual during each event is determined based on the Controlled Vocabulary (e.g., Free, Enslaved, Liberated, Fugitive). The age of the individual at the time of particular event(s) is recorded as found in the primary document or otherwise imputed or associated with a range of years. Age categories, including child, adult, elder, are determined at the time of a specific event according to the Controlled Vocabulary. The start date and end date for an “Event” (year-month-day) is recorded according to documentation or imputed when the date is not cited in the original source but has been determined by calculation. An example of calculating an imputed date would be finding out the year when the “birth” event took place by subtracting the known or approximate age of an individual from the year the information was recorded or the event occurred. The circa period is clearly noted when the date is an approximation (a range of years). Religious affiliation at the time of the event is noted, including reference to a specific brotherhood, church, ritual society or other association if it can be determined.
The event description is a summary that captures a single event to clarify the purpose of the event or if something particular happened during an event. For example, a “resistance” event could mean the individual ran away from the master, started a rebellion, or refused to work and hence requires clarification. A “birth” event, by itself, is usually clear enough. However, if the primary document states that there were complications during the “birth” event, then details are noted. An event identifier provides a unique reference for each event to recognize connections among more than one person to the same event, making it possible to search a specific event and everyone who is known to have participated on it. Events are connected to specific sources on the portal, as is all other data. The occupation of the person is noted for each event based on the Controlled Vocabulary. If other individuals, not including relatives, are involved in the event, reference is made to them based on the Controlled Vocabulary. For example, for an enslaved African the ship captain in the “Departure” and “Arrival” events can be determined through the Slave Voyages database; ex. Captain | Henry Thorne. Ideally, the precise location where the event took place is noted, if known. If no precise location is given, a specific area (Street, City, District, Country, Region) is noted. The place where an event occurred is referred to according the place type, ranging from streets to regions, with the exact coordinates of the location. Variant spellings of the names of places in primary sources and modern equivalents are noted.
Details on the sources used in reconstructing the known history of individual are standardized so that users can access the exact passages relevant to each individual and the full document, including the complete name and details of the original documentation. If the original source does not have a title, the ID number of the document is entered, the source it was taken from, and the date it was published or otherwise identified. The origin of the digital source is recorded, along with the name of a person or organization that made a significant intellectual contribution to the resource. Also included are the name of the institution that possesses the original collection, the exact date when the original document was created or published, the language in which the original document was written, date at which the digital resource was created, and the type of the resource, such as text, iconography, picture, and the electronic format in which the digital resource is made available (PDF, JPEG). Also entered is information about rights held in and over the resource and a consistent reference to all institutions or administrative unit(s) that contributed to the creation, management, description, and/or dissemination of the digital resource. A general term is used that describes what the original resource is or the format that it takes, such as slave advertisement, newspaper article, book. The number of pages in the document or source is recorded.
Best Practices relies on a methodology that uses Controlled Vocabularies, the fundamental building blocks by which is meant the terms that are used to describe and present data that are carefully defined and applied in the organization of data. Since words and symbols have different meanings in spoken and written language than in a programming language, with particular attention to clear definitions and use vocabularies that are exact and consistent, which are developed as “Controlled Vocabularies.” Otherwise data are not “clean” and cannot be processed by computer algorithms. Controlled Vocabularies are used to enable the development of a database that allows analysis of patterns and facilitates interpretation. In some cases, such as in identifying gender, the Controlled Vocabulary is simple (male, female, unknown). In other cases, the Controlled Vocabulary is relatively straight forward, such as the Controlled Vocabulary for Physical Descriptions (scarification, tattooed, missing limbs, slave branding, yellowish complexion, dark complexion, Poro marks, partially blind, totally blind, smallpox marks, marks of illness). Other Controlled Vocabulary relating to ethnonym, occupation and other details are more complicated and expand as new identifications are found.
The database identifies individuals through a unique identification number FN000000 that allows a potential database of a million people. For purposes of reference, a single name is taken as a primary name, which can be the name as reported in a source, a name that is the result of conversion to Christianity, or a nickname that was used for that person. All alternative names are recorded, so that the database can be searched for any variation. If a name is changed, such as through conversion to Christianity, the change is recorded as an event. Freedom Narratives does not privilege the use of surnames, which is an approach adopted by some databases and the Enslaved hub. We consider the emphasis on surnames to be Eurocentric and reflect a Judeo-Christian tradition and not traditions that derive from the societies of the enslaved themselves. As noted in our distinction between “slave narratives” and “freedom narratives” makes clear, there were two categories of enslaved individuals, first those who were born in Africa and second those who were born in the Americas. Controlled Vocabularies that use a category of identification that is referred to as “surname,” which is most often a Christian nomenclature, and occasionally Jewish, are irrelevant to the vast majority of cases involving enslaved individuals. The identification of surnames may be of use in some cases, admittedly, as in the fact that the enslaved who were able to gain their freedom in the Americas could take the surname of their master/mistress, and in cases of fugitive slaves such reference to individuals through the use of the owner’s name was a means of identification. Baptism in Portuguese and Spanish colonial settings often involved the adoption of surnames, with some variation from practice among the free population. Spanish practice among the free population used the surnames of both the father and mother, with the father’s surname usually coming first. Portuguese nomenclature could refer to other social relationships, not just the preferred surname of the master.
The enslaved who were born in Africa seldom if ever had surnames, at least not before conversion to Christianity. The use of surnames ignores the Muslim component of the enslaved population, wherein Muslims did not use surnames, and it conflates naming practices into an indistinguishable pattern. The insistence on surnames also mistakenly confuses African naming traditions and can incorrectly refer to someone like the individual whose birth name was Olaudah Equiano by the supposed surname, Equiano, when Equiano was not a surname. For Freedom Narratives, the concept of surname is not used. Instead, individuals are identified by a unique identification number (FN000000) and the most common name used in the sources. Variations of names, nicknames, Christian names, variations in spellings are all included under alternate names separated by a semi colon. If the only name that is known is a Christian name, then first and last names are included together, when known, separated by an underscore.
For those who were born into slavery in the Americas or elsewhere other than West Africa, a brief reflection is warranted, even though this is not generally issue for the Freedom Narratives database, unless someone was born outside of West Africa and returned to West Africa and hence comes to the attention of researchers. Individuals in the Americas were given slave names that usually lacked a surname or other designation except in relation to their master/mistress. Such identification is important, of course, and have to be noted, just as Christian names taken at baptism have to.
Names can provide clues to identity when they can be correlated with other factors, such as religion in the case of Muslim names, with days of the week, as in the case of Akan day names, and specific cultures when the meanings of names in a language can be discerned, as in the case of Olaudah Equiano. Attempts to equate names with specific ethnic groups are less certain, despite the attempt of to do so with reference to individuals taken off slave ships by the British Navy in the nineteenth century. With respect to Freedom Narratives, this approach has very limited veracity. Names taken alone can be misleading because of possible similarities in the pronunciation of names that are not otherwise known to be identical in different societies and the uneven filtering of phonetic rendering of names by third parties who do not know the languages and cultural practices of those being attributed with specific origins. The difficulty of relying on names to establish identity is particularly pronounced when a correlation with ethnicity is claimed.
One difficulty of what is known when ethnicity is assumed or assigned relates to the meaning of ethnicity, which in itself is contentious. We argue that references to ethnicity, tribe, etc. usually refer to language. Moreover, even when language is inferred from the reference, it is usually made in a manner that suggests that the individual knew only one language when in fact in most cases we know individuals spoke at least two and often several languages. We lump ethnonyms and language together because it is not usually possible to understand what a source is referring to when ethnicity is claimed. The second problem is that the meaning of ethnicity in the diaspora in the Americas is not equivalent to ethnic designations in Africa. In the Americas, various terms are usually described as “nation” or “country,” whether in Portuguese, Spanish, French or English, which are vague constructs that do not correspond with designations referred to as ethnicity in Africa. Second, “ethnic” terms in the Americas, such as Lucumí in Cuba, Mandingo/Mandinga, Angola, Congo are broad categories that do not correspond to terms in Africa, despite fuzzy links. The possibility exists that individuals were interviewed or responded in Yoruba whether or not they actually also spoke another language and Yoruba was not their mother tongue. When the term is hyphenated with some other designation, such as Hausa, all that can be assumed is that the person so identified spoke at least two languages, Yoruba and Hausa.