Using the Full Databases
THE LOUISIANA SLAVE DATABASE AND THE LOUISIANA FREE DATABASE:
By Gwendolyn Midlo Hall
While the search engine supplied here is a good way to get started, the creation, application and utility of these databases is a big subject and requires detailed explanations.
These databases and supporting documentation, calculations and images were first published in Gwendolyn Midlo Hall, ed., Databases for the Study of Afro-Louisiana History and Genealogy 1699-1860 Baton Rouge: Louisiana State University Press, 2000). This was a compact disk publication that also included census databases and spreadsheets by Paul LaChance and by Jeffrey and Virginia Gould. Final design, manufacture, and packaging of this compact disk publication was very ably overseen by Joel Williams and carried out by Tod Love and others at the Computer Services Center of Louisiana State University. Maureen Hewitt, Associate Director and Editor-in-Chief of Louisiana State University Press, threw herself into this project with as much enthusiasm and dedication as I did, carefully, skillfully, and tactfully sponsoring and editing the final product.
The creation of these databases began in 1984 after I discovered an impressive number of notarial documents in the courthouse at New Roads, Pointe Coupee Parish, which indicated the African ethnicity of the slaves. Robert A. Rosenberg, Director of the Edison Papers at Rutgers University, helped, encouraged, and taught me how to use databases. Results and calculations for Pointe Coupee during the Spanish period went into my book, Africans in Colonial Louisiana: The Development of Afro-Creole Culture in the Eighteenth Century (Baton Rouge: Louisiana State University Press, 1992).
These databases were made possible because of the generous support of the National Endowment for the Humanities and other funding agencies supplying matching funds and other funding. In 1991, the National Endowment for the Humanities funded a contract mainly to expand the databases that I began in 1984 in space and time. All matching funds allocated by the National Endowment for the Humanities were met by contributions from the Cultural Affairs Office of France, the Program for Cultural Cooperation between the Spanish Ministry of Culture and United States Universities, the John Simon Guggenheim Foundation, the Louisiana Endowment for the Humanities, the Historic New Orleans Collection, and the University of New Orleans, with in-kind support from Northeastern University. Aside from administering the grant, Rutgers University did not collect overhead costs during the first two years of the project and provided very generous released time for me in return for a very modest compensation from the NEH contract.
The project finally included all the geographic areas now constituting the State of Louisiana through 1820 and also included some documents originating in or involving parts of what are now Mississippi, Alabama, and Florida. The vast majority of the documents used are housed in Louisiana, but a few were consulted in archives in France and Spain and at the University of Texas in Austin. The project stopped with 1820 when descriptions of the African ethnicities of slaves became sparse in Louisiana at the same time as the volume of documents escalated.
I was project director and co-principal investigator with Patrick Manning. We were very fortunate that database and statistical software had greatly improved since my original databases were created, so that by 1991 we were able to define many more fields and input many more records than was possible when my project began in 1984. Patrick Manning contributed to the expansion of the database definitions and the creation of code sheets when our project began in 1991. It is hard to overestimate his role in avoiding mistakes that would have haunted us ever after. Thereafter Manning devoted most of his research time funded under this project to developing and researching his Louisiana Sacramental Records Spreadsheets which will be published in the future. Just as the project neared completion, compact disk writers became available for home computers and made it possible for me to do just about all of the technical work myself.
The choice of dBASE V for DOS as the software package for the expansion of my original databases was a good one. The data are entirely portable and can be uploaded into new database and statistical software packages for Windows or Apple and beyond as this technology rapidly advances. The data can be uploaded into good database or spreadsheet software, recoded, queried, reorganized, and/or uploaded into SPSS (Statistical Package for Social Sciences) or other statistical software to allow for easy recoding, complex calculations, and for making graphs. In the process, the data can be reorganized or redefined to suit the user. Complex computer programming has been avoided, keeping costs down and flexibility up. My philosophy is that the conceptualization and design of historical databases are the work of historians, not technicians. Compact disk writers for desktop computers appeared on the market just as publication and manufacturing of the databases began. As a result, much time, money, and aggravation were saved and we were able to keep the price of the compact disk published by Louisiana State University Press in 2000 at only $45. I am happy to see that my complete databases as well as the software to use them are now available through ibiblio.org free of charge.
Database design cannot be divorced from the process of research. Early database design does not produce a finished product. It evolves as the historian becomes more familiar with the documents, with the data, and with new questions as they arise. Answers, tentative answers, and partial answers lead to new questions. Additional fields are entered and recoded during and after the data entry process has been completed. Maintaining flexibility to make changes in database design and to make corrections in data entry is essential. Once we largely completed the data entry process, many details had to be ironed out--always a slow, tedious process on a large database. Checking for duplications of identical documents found in more than one archive and deleting the duplications was a major chore. We did not delete the same slaves duplicated over time because, aside from the difficulties and uncertainty of such an undertaking, it would have eliminated much data, which demonstrated changing patterns and made it impossible to trace particular slaves in time depth. When estates were inventoried and then sold immediately afterwards, we recorded slaves only once, indicating the date of the inventory and the date of the sale and both the inventory and the sale price. In cases where months and, at times, years elapsed between inventory and sale of estates, we recorded slaves separately in order to be able to calculate changes over time in the same estate.
Patrick Manning played a major in expanding the identification of African ethnicities and their African regions of origin at the early stages of the NEH project. Beginning in 1996, members of the international community of scholars generously contributed their expertise in identifying still unidentified African ethnicities over H-Africa.net and through email. The input of Adam Jones, Owen Kalinga, Martin Klein, Robin Law, Michael Levin, Paul Lovejoy, Joseph C. Miller, Bruce Mouser, Mikael Parkvall, and Stephen Rockel is gratefully acknowledged. Dr. Ibrahima Seck, an outstanding historian living in Senegal who is of mixed Wolof and Fulbe ancestry and has a strong feeling for language, helped identify many of the African names (Code 4 in NAMETYPE fieldabout 10,366 names), names which were possibly African or European (Code 8 in NAMETYPE field-about 8,743 names), and Islamic names (about 1,017 names). He also included explanations of some of these names in the NAMEXPLAIN field where his input is initialed IS.
Pinpointing and identifying some of the African "nations" (ethnicities) found in the documents was more difficult than one might expect. Much information was initially taken from Philip D. Curtin's The Atlantic Slave Trade: A Census. But an amazing array of varieties of "nations" was found in the documents. There are about 9,000 records that contain specific African ethnicity information. These involve 217 different ethnicities among which 96 have been identified and 121 (consisting of only 152 individuals) whose "nation" was given but remains unidentified. All of these unidentified ethnicities are represented by very few individuals. Among the 8,840 Africans of identified ethnicities, 8,508 (96.2%) were clustered among 18 ethnicities who have been recoded in SPSS as AFREQ. This recoded field ranges between a low of 68 records for the Edo of the Bight of Benin to a high of 3,035 for the Congo of West Central Africa. Although Africans represented by very few ethnicities are of interest to specialists in African history, they are too few in number to use for studies of distribution of ethnicities in Louisiana.
A number of people assisted with the data entry carried out in courthouses and archives throughout Louisiana between July 1, 1991 and August 31, 1996. Ulysses S. Ricard, Jr., Chief Archivist of the Amistad Research Center, received released time to work full time as associate researcher on the project during its first year. Unfortunately, he became quite ill at the outset of the project, emerged from the hospital nearly blind, and died two years later. Ricard's enthusiasm for the work and his skills in the paleography of Louisiana French and Spanish colonial documents were an inspiration to all of us. He is sorely missed. This publication is dedicated to his memory. Philip S. MacLeod, then a student in Tulane University's Latin American Studies Ph. D. program, filled in when Ricard got sick and worked full time on the project during the first year, continued to work at least part time throughout the entire project and devoted much time and care to the final checking, translation into English of skills and illnesses, and cleaning of the databases before publication. His diligence, skill, modesty, dedication, open-mindedness, cares, and attention to details in research and data entry was invaluable. Despite the years he devoted to this project, he was able to complete a fine Ph. D. thesis about colonial Costa Rica. He was one of the last Richard L. Greenleaf student to complete his Ph. D. He is now Archivist for Manuscripts and Archives Sterling Memorial Library Yale University. Gregory Osborne moved to New Orleans from Los Angeles to work full time on the project during academic year 1991-92. He now works in the Louisiana Room of the New Orleans Public Library and devotes much of his time and attention to Afro-Louisiana genealogy. Mabel Macias devoted careful time and attention to research and data entry for about a year. Peter Caron and Osvaldo Ortega each entered data for a few weeks. Liliane Chauleau, Director of the Archives of Martinique, now retired, made several trips to New Orleans and then consulted the microfilms of the Louisiana Superior Council Records, assisting with some of the especially thorny problems of paleography of eighteenth century French documents. Mabel Robinson Williams helped me reorganize some of the fields and checked for typos in the name fields in preparation for publication. Inevitably, there are still some errors on such a huge database, but I hope they are minimal.
This project owes a deep debt to Paul Lachance who spent a full week showing me how to use SPSS for Windows during the summer of 1996 and created the initial SPSS syntax file for the Louisiana Slave Database. This has allowed me to make calculations and graphs on my databases without assistance from others. His help is gratefully acknowledged. SPSS.syn (syntax) and SPSS.sav (data) as well as sample graphs calculated from these files are included on these pages. The SPSS syntax files allow the user to customize the databases in accordance with the user's opinions and needs. The SPSS syntax files provided on this disk contain value labels for each field, which translate digits into words. JNLSLAVE.syn recodes birthplaces of Africans by coastal region under the field AFREGION. The recoded field AFETH disaggregates Africans assigned to various coastal regions based upon "nation" descriptions in the documents from those identified by coastal regions only. Slaves described as "Guinea" or from the "Coast of Guinea" were placed under the Sierra Leone coastal region based upon the author's study of the context in which these terms were used. They can be easily moved to the Africans of unknown origin category by changing one digit on the syntax file if the researcher so wishes.
These databases were created almost entirely from original, manuscript documents located in courthouses and historical archives throughout the State of Louisiana. The project lasted 15 years but was funded only between 1991 and 1996. I retired from Rutgers University in 1996 in order to complete the project, which had expanded beyond my wildest expectations because of the extraordinary quantity and quality of documents found. All sources and repositories where these documents were found are fully cited in the DEPOT field. There is complete document retrieval information for every record (individual slave) contained in these databases. Some records were entered from original manuscript documents housed in archives in France, in Spain, in Cuba and at the University of Texas in Austin. A few were entered from transcriptions of original documents published in books and journals. Some were supplied by Emily Clark from her research in the archives of the Ursuline Convent in New Orleans. I entered records from French Atlantic Slave Trade Voyages from original manuscript sources in France and in Louisiana. Records for British Atlantic slave trade voyages to Louisiana were entered from a pre-publication version of David Eltis, David Richardson, Stephen Behrendt, and Herbert S. Klein editors, The Transatlantic Slave Trade: A Database on CD-ROM (Cambridge, 1999).
In databases created from original manuscript documents, fields are defined which can answer the questions their creators wish to pose. Definition of fields is a long process requiring a thorough knowledge of the documents used. As the researcher becomes more familiar with the documents, fields are added as regular significant information appears in the documents. There is a COMMENTS field in these Louisiana databases where important information for which there were no defined fields were placed and from which new fields were created. Complete descriptions and explanations of the fields of each database are contained in the files SLAVECODE and FREECODE.
Data were almost always entered directly from the original manuscript documents into dBASE V for DOS using laptop computers. I entered data alongside the other researchers while training and supervising them until I was confident of their competence and then carefully reviewed their work, correcting detectable errors and maintained the central databases into which new data were downloaded as they were collected. These databases contain at least 1,000,000 entries of often-complex information. Standardized spellings did not exist in these documents. Names were entered the way they were spelled, and spellings of the same names varied among documents. They were written in three languages: French (60,914 or 60.5%), Spanish (17,848 or 17.7%), and English (21,878 or 21.7%). We had to identify and translate words for skills and illnesses some of which are no longer used in any language. In the process of translation, we made every attempt to include the original words and spellings in the original languages in all fields as well as in phrases and sentences in the COMMENTS field. There is a LINGUISTICS field that flags records relevant to linguists. Most of the documents were located in Orleans Parish (58,058 or 57.7%), the rest (42,582 or 42.3%) in rural Louisiana. In rural Louisiana throughout the Spanish period, almost all documents were in French. By 1804, after the United States took over Louisiana, the rural documents became heavily English while the New Orleans documents, which had been almost entirely in Spanish during the Spanish period (1770-1803) reverted largely to French with substantial numbers in English as well.
Each record in these databases describes an individual slave. Throughout the Americas, slaves were listed as property and therefore described and inventoried after the master died, when slaves were bought and sold, when they were reported as runaways, when slaves testified in court cases including interrogations about running away, conspiracies and uprisings against slavery, and in several other types of documents including wills, marriage contracts, leases, seizures for debt or for the criminal activities of the master, mortgages of slaves, and reports of death. The Louisiana Slave Database also includes all clearly documented voyages that brought enslaved Africans from various African coasts to Louisiana. It contains well over 100,000 records (individual slaves) and 114 fields (bits of comparable information). Complete document retrieval information is included in each record. Type of document, name, birthplace, numeric age or age category, gender, racial designation, if the slave was involved in running away/and or in conspiracies and revolts against slavery, family relationships from which child/woman ratios and age of mother at the birth of her first surviving child and age specific marriage patterns can be calculated. Skills, illnesses, character, African language group, language of document, and linguistics are some of the fields. The VIA and the WENT fields record where the sellers came from and where the slaves were sold to when this information was given, regardless of the slave's birthplace. Two fields FREE and ESTATEFREE indicate if there was someone described as at least partially of African descent involved in any of these transactions as buyer, seller, or the deceased. I collected documents recording the same price in different currencies over many years. Robert A. Rosenberg, Director of the Edison Papers Project at Rutgers University, helped me develop the price calculation formulas in 1988 using comparable price data collected from the Pointe Coupee Post during the Spanish period. The original formula has held up remarkably well as comparable prices continued to be collected for the entire range of times and places represented in these two databases and the original price conversion formula was used. Both inventory and sale prices indicated the currency used as well as the amount paid in this currency. Two fields, INVVALP and SALEVALP convert both prices into common denominator prices. A few changes were made in price fields record by record all of which are indicated in the comments field. Subtleties still need to be ironed out, but the price data are abundant and largely reliable after 1769. Earlier prices, especially between 1735 and 1769, were quite unstable due to price inflation. The GROUP field indicates if the price or prices is of an individual, of a group, or if no price information exists at all. Only individual prices must be chosen for price calculations at this time, but group prices can eventually be recoded to allow for calculations.
The Louisiana Free Database (currently available in the downloads section of this site only) contains 4,068 records and 62 fields involving manumissions of slaves. The fields include MEANS, the means of manumission, FREEREL, relationship of freer to freed when a third party purchased the slave in order to manumit, REASONS given for manumission, GRATUITOUS, if no payment was made, CONDITIONS imposed, for example, additional time of service required, SALEVALP, the price paid in case of a purchase or self-purchase, FREER, the name of a third party involved in the manumission, and FREED, the likelihood that the slave was actually manumitted. This database contains family relationships among the manumitted including names, racial categories, ages, and birthplaces of mothers of slaves involved in manumissions. WHITEDAD indicates if the manumitted was certainly, probably or possibly freed by a white father. There are first and last names, gender, racial designations, and birthplaces of the deceased, the masters, sellers, buyers, and freers in these two databases.
These databases will be very useful to genealogists. It is likely that most people and/or families who lived in Louisiana through the year 1820 regardless of racial designation or status can be found in these databases by name. Many people owned slaves in early Louisiana. Aside from large slaveholders, many whites and free people of African descent owned at least one or two slaves. They were sometimes domestic slaves and some of them were blood relatives of their owners. Thus, a very broad range of people who lived in Louisiana through 1820 is recorded in these databases. The names of the slaves can be linked to the names of the masters as well as their own family members over time. Precise information recorded in every record includes the exact date, location and the document number when one exists, and the name of the notary when relevant. This information will allow the genealogist or scholar to easily find and copy the original document.
Those searching for names must use their imagination about how names could have been spelled. Both SOUNDSLIKE and name indexes should be used and various spellings tried. While the search engine allows for a SOUNDSLIKE search, there are some names written in different languages which will not be picked up: for example, Santiago in Spanish would be written Jacques in French. The best way to search for all possible spellings of names would be to create an alphabetical index of the relevant name field and search for every conceivable spelling.
Correction of typos found among hundreds of thousands of names and recoding and reorganization of some fields for ease in calculations and the creation of graphs took place in the final stages of database preparation. Errors had to be checked and corrected. Corrections were made with maximum respect for the data since standardized spelling was not the norm and names spelled differently could have been distinct names.
Some of these data have been entered into the fields in the form of digits. Because digits are more precise than words, they generally improve the accuracy and uniformity of the data and facilitate complex, statistical calculations. An explanation of what each digit represents for each field is contained in the files SLAVECODE the code sheet for the Louisiana Slave Database and FREECODE, the code sheet for the Louisiana Free Database, but it is rarely necessary to consult them. Genealogists searching for ancestors using database software alone have very few digits to deal with. The name fields are not digitalized. The genealogist can do a name search, preferably by using the alphabetized index of their software and locating a particular early Louisiana person or family, pinpointing where and when they lived or died, and obtain precise information about where these documents listing their names are housed so they can be easily located and copied if desired. It is sometimes possible to identify the ethnicities or "nations" of the African ancestors of individuals who have ancestors in Louisiana by making links between information contained in the Sacramental Records of the various Archdioceses of the Catholic Church of Louisiana and these databases. The vast majority (74.4%) of ages are recorded in the Louisiana Slave Database in numeric form. A much smaller percent (15.3%) give age categories, which are digitalized. Skills, illnesses, character as described by the master or appraiser, family relationships, are all described first in plain English and then digitalized. The COMMENTS fields have been translated into English. The various currencies in circulation were converted into common denominator prices. Parish names are abbreviated in English in the PARISH field and should be clear to the user without consulting the digitalized LOCATION field. Many of the African birthplaces (meaning ethnicities) as well as other birthplaces are written out in the SPELL fields and record the exact way the "nation" was spelled in the document, but the codes for the BIRTHPL fields will sometimes have to be looked up in the Codesheet files by researchers who do not use SPSS or other statistical software which automatically converts the digits into words.
For researchers making complex calculations on SPSS Version 9 or higher, the SPSS.sav files supplied on this site for download automatically translate all digits into words. When the researcher needs to know what these digits represent in order to select data for calculations, the researcher simply clicks on the icon for codes which looks like a vertical ruler on the menu line and then clicks on the relevant field and the SPSS package displays value labels for each digit of the field. SPSS syntax files are supplied for both databases. These syntax files can be altered to recode and reorganize the data in any way the researcher wishes. (To learn more about this excellent software, visit the SPSS website.)
These remarkable Louisiana documents pay particular attention to recording the birthplace of the slaves described, including their African "nations." The richness of this information is certainly unique for any documents that describe slaves who became part of the population of the United States. The vast majority of slaves whose birthplaces were identified were Africans. Among 38,019 slaves whose birthplaces were recorded, 24,349 (64.0%) were of African birth. Among these, 8,994 (36.9%) indicate specific "nations" (ethnicities), 9,382 (38.5%) indicate their African coastal origins only. The vast majority of the latter involve Atlantic slave trade voyages (which can be excluded from calculations in queries by selecting doctype<25). Some 5,973 records (25.3%) simply indicate that they were Africans with no other information about their origins. Information about the maritime slave trade is rich: 11,139 records contain information about the voyage individual slaves arrived on, among whom 8,655 arrived on Atlantic slave trade voyages and 2,484 on transshipment voyages from the Caribbean and from East Coast ports of the United States. The vast majority of slaves of identified origins transshipped from the Caribbean were newly arriving Africans (listed as brut in French or bozal in Spanish) purchased from Atlantic slave trade ships. There is a particularly high percentage of identified birthplaces, especially many African "nations" of slaves, recorded in documents dating between 1770 and 1820 in lower Mississippi Valley parishes: St. Charles, St. John the Baptist, Pointe Coupee, and to a lesser extent Orleans. This remarkably full information allows us to make accurate estimates of the proportions of Creoles, Africans of various "nations," Caribbean, English-speaking, and Indian slaves over time. There are 1,069 records involving runaway slaves (select maroon=1) and 589 records which indicate and record some testimony by slaves (select doctype=24). To access all records describing slaves involved conspiracies and revolts against slavery, select revolts=1. There are 22,274 records (22.4%) which show family relationships among slaves. At least one character trait is shown for 855, at least one skill for 8,745, and at least one illness for 2,392 slaves. Complete inventories for 2,904 individual estates are recorded with estate numbers that show, in addition to a unique number, the parish, the number of slaves on the estate, and the year.
In short, the Louisiana Slave and the Louisiana Free Database are unique tools for studying the slave as well as the free population of Louisiana through 1820. They contain a vast amount of rich and unusual, hitherto unexplored and unrecorded data and organize them to ask and to quickly and easily answer many complex, previously asked and unasked questions. They should be helpful to scholars making comparisons of patterns of slavery in other places in the Americas. They should help clarify patterns of the slave trade in Africa as well, pinpointing which ethnicities had entered the Atlantic slave trade by a certain time, their gender, the changing proportions among ethnicities taken from various African coasts over time, their marriage and parenting patterns, their resistance to slavery, and their age patterns. They will certainly be very useful to anyone searching for a Louisiana ancestor.
I demonstrate the usefulness of these databases for historical research and interpretation in several recently published and forthcoming articles: "In search of the Invisible Senegambians : the Louisiana Slave Database (1723-1820)", in Saint-Louis et l'Esclavage, Actes du symposium international sur la traite négrière à Saint-Louis du Sénégal et dans son arrière-pays, Saint-Louis, 18, 19 et 20 décembre 1998, Djibril Samb Editor, IFAN (Institut Fondamental dAfrique Noir, Dakar, Senegal) , Université Cheikh Anta Diop de Dakar, Initiations et Etudes Africaines N° 39; "The Meanings of Mina," in Paul Lovejoy and David Trotman, editors, Trans-Atlantic Dimensions of Ethnicity in the African Diaspora (London and New York: Continuum Press, forthcoming June 2002); in a book I am writing tentatively entitled The African Diaspora in the Americas: Regions, Ethnicities, and Cultures; and in another book, Ethnicity and Race, Slavery and Freedom: a Comparison of French, Spanish, and Early American Louisiana (Chapel Hill: University of North Carolina Press, forthcoming) . The Louisiana Slave Database has been used to help solve a thorny problem in creole linguistics in Thomas A. Klingler, "Louisiana Creole: The Multiple-Geneses Hypothesis Reconsidered," Journal of Pidgin and Creole Linguistics, summer 2000. The Louisiana Slave Database has turned out to be of keen interest to geneticists as well.
It is my hope that these databases will help other scholars and genealogists in their fascinating and absorbing work. Best of luck to all of you.