A Note on David Herlihy's Computer Files.

R. Burr Litchfield, Brown University.

From his first book Pisa in the Early Renaissance (1958), David Herlihy applied a method of historical research that was well suited to computer analysis. "Serial History" (so-called by the French Annalistes) examines series of archival documents that provide uniform information about individuals over time--notarial registers, cadasteral tax lists, records of births, marriages, and deaths, registrations of office-holders. The systematic analysis of these records permits assessment of individuals in groups and the delineation of social structures and collective destinies. David did not make use of a computer in his book on Pisa, or in his second book, Medieval and Renaissance Pistoia (1967). At that time even the most advanced historians seldom went much beyond using punch-cards and a card sorter.

Soon after his move to the University of Wisconsin in 1964 David blossomed as an enthusiastic computer user. At that time he conceived the project for analyzing the entire Florentine Catasto of 1427-29, the surviving archival records for the tax on wealth imposed by the Republic of Florence on the city of Florence and its other territories in Tuscany. Fernand Braudel of the Paris Ecole Pratique des Hautes Etudes offered support and the collaboration of Christiane Klapisch-Zuber for this gigantic enterprise. Subsequently, at Harvard and at Brown University, David continued to work methodically, up to the year of his death in 1991 when he was President of the American Historical Association, adding to and analyzing the computer data files for the Catasto and other sources that underlay his later works. [1]

Anyone who uses a computer for historical research knows that accumulated data can be useful to other scholars after publication of the initial works derived from them. The Florentine Catasto is one of the most basic and important documents for Florentine social history during the Renaissance because it provides significant information about most inhabitants of the city. The purpose of this note is to provide some insight into David's coding of the Catasto, that is reflected in the Brief Codebook of the 1427-29 Catasto Data File for Florence included with this World Wide Web page. The materials from the Florence Archivio di Stato that form the basis of the Catasto data file are fully described in Les Toscans et leurs Familles, and more summarily in the English abridged version of this work: Tuscans and their Families. Anyone making use of data from the Catasto should study these works carefully. The Brief Codebook shows that the data included in the Online Catasto are abridged from the complete Catasto data set used in Les Toscans et leurs Familles, and even the complete set included only part of the information available in the original records of the Catasto in the Florence Archivio di Stato. The Online Catasto provides a summary of wealth of each taxable household in the city, with names and information about the head of household, and the total number of household members (bocche). It reflects the initial tax declarations of 1427 plus additions and adjustments made in 1428 and 1429. The data is arranged so that one can search the data file by family name, first name, and patronymic of household heads. The complete Catasto data set used in Les Toscans et leurs Familles provided information for both the city of Florence and all other cities and rural districts under Florentine rule in Tuscany. It also included the sex, age, marital status, relation to head of household for each subordinate household member (although the names of subordinate household members were not included). The information for subordinate household members was omitted from the Online Catasto because of the difficulty in managing the non-rectangular data format in which the data for these individuals was initially coded. But this information can be obtained from the version of the complete Catasto data file distributed by the Data and Program Library Service of the University of Wisconsin [2] The original documents of the Catasto, the volumes of Campioni in the Archivio di Stato in Florence, of course, contain much additional information that was never coded by David Herlihy and Christiane Klapisch-Zuber. These documents provide names of all individuals, as well as more detailed information about the wealth of households, and other information. The volume and page numbers of individual accounts in the Campioni are provided in the Online Catasto, however, to help researchers use the original documents. Microfilms of the Catasto are available not only in Florence, but also at the University of Wisconsin Memorial Library in Madison Wisconsin, the Center for Research Libraries in Chicago, and Widener Library at Harvard University in Cambridge, Massachusetts. [3] Researchers thus have access both to usable versions of the computer data files of the Catasto and to microfilms of the original documents on which the computer files were based.

Users of the Online Catasto should be aware of some particular characteristics of the data file. The Catasto is an historical document with many eccentricities, and coding the data for computer analysis introduced further problems of interpretation. Most of these problems are discussed in Les Toscans et leurs familles. You should remember that your World Wide Web browser will probably allow you to print out documentation from the Online Catasto for your own reference, particularly the Brief Codebook and the List of Family Names Found in the Catasto. Two general types of considerations should be kept in mind in your initial exploration of the data:

1. The Series variable makes some difference. This distinguishes whether the entry was one of the original declarations of wealth made in 1427 (9,780 households in Florence) or was one of the entries added or modified in 1428 (101 households) or 1429 (123 households). If you want a snapshot cross-section of Florence you should limit your search to 1427 (Series=1). The 1428 (Series=2) and 1429 (Series=7) data sometimes duplicates, or simply corrects, the initial entry. For instance, Filippo di Tommaso Alberti submitted his tax declaration three times, with minor adjustments in the value of his property, in each of the three years. The Series variable also impinges on the size of the population of Florence. Here one should clearly limit oneself to the declarations of 1427, but caution is needed. By summing up the total of the Bocche variable, for instance, one would arrive at a total population for Florence in 1427 of 38,269 individuals, while Les Toscans et leurs Familles (p. 183) gives the population in that year to have been 37,144. The discrepancy appears to have resulted from the judicious exclusion from the total by David Herlihy and Christiane Klapisch-Zuber of babies born after 1427, who were added to the original 1427 declarations rather than to the later years, and also of individuals who were absent from Florence for reasons of commerce, etc., although their property was assessed in the Catasto. When these exclusions are taken into consideration the two population totals coincide.

2. One needs to be aware that there are some inconsistencies in interpreting the original documents of the Catasto in the data file, and even some miscoding of data. For instance, family names, names, and patronymics were not fully standardized in the coding and were truncated beyond ten letters. (They will be unrecognizable to the SQL searching program unless they are typed in exactly as they appear, and in CAPITAL letters.) To make it easier to identify family names we have appended a List of Family Names Found in the Catasto, for the 37 per cent of households that had family names. Even among these it is clear that some family names are not surnames, but rather approximations of occupations, or other comments. For instance, the first name appearing in the list of family names, "ACCANTASAN", appears to refer to someone living next to some church. The second name, "ACCATTAFAN", appears to indicate that the individual was a beggar. One notices from the third and fourth entries that the family name generally appearing in later periods of Florentine history as "ACCIAIUOLI" appeared in the Catasto with two variant spellings: "ACCIAIOLI" (7 cases) and "ACCIAUOLI" (1 case). The search screens, however, permit you to do a combined search for the two surnames (Family_name = ACCIAIOLI [or] Family_name = ACCIAUOLI).  One would search in vain among the truncated family names reading "BRUNELLESC" for the architect Filippo Brunelleschi. But searching among individuals with the Trade variable coded 19 (Miscellaneous), through serendipity, one indeed finds an individual whose name was coded from the Catasto as "FILIPPO", and whose patronymic was recorded as "BRUNELLESC"--the architect we were looking for. It may be that his name appeared in the Catasto in this way. Similarly, the ancestors of Michelangelo Buonarroti, do not appear in the Catasto under "BUONARROTI" (a family name missing from the Catasto), but instead under the family name "SIMONI" (which, indeed, was their family name at the time). The name of Cosimo de Medici does not appear in this version of the 1427 Catasto. This is because Cosimo was listed as a subordinate member of his father's household (Giovanni di Bicci de Medici), and thus he is counted among the 10 Bocche in Giovanni di Bicci's household. Other problems emerge with occupations, the Trade variable, where many individual occupational titles were grouped together under a few numerical codes. With regard to wealth, the Taxable (taxable assets- - that is: total assets minus deductions) of the richest man in Florence, Palla di Noferi Strozzi, is given in the data file as 99,999 florin, rather than as the 103,306 florin that would appear logically from other items in his account to have been the true figure, probably because only five digits were available to code deductions--and Palla di Noferi Strozzi was the only person in the city to have assets that reached such a large figure.

We have not made any effort to correct such inconsistencies in the Online Catasto data file, but we indicate their existence so as to caution users to exercise care and ingenuity in their research. Only recourse to the original registers of the Catasto, with the help of the volume and page numbers provided in the Online Catasto, would permit one to definitively resolve such uncertain cases.