The Ragusa Project began as an attempt to explore some of the interconnections between social anthropology and social history. It had two major aims, one substantive and the other methodological. The substantive aim was to investigate the political and economic careers of the members of an entire ruling class, the patrician e\*'lite of a fifteenth-century city-state. I hoped to investigate the interplay between individual careers in politics and commerce and the interests of larger groupings such as lineage, class, and state to which these individuals belonged.
The methodological goal was to test the feasibility of using very rich archival sources, and the data processing capabilities of the computer, to `reconstruct' a total governing class. In this paper I describe the data collected during the project, the database system which stores it on computer, and some of the ways that information are being extracted.
Several features of Ragusa (Dubrovnik of the 15th century) make it especially suitable for this type of study. First, the metropolitan centre was small in comparison with comparable units. It had a population of about 6,000 inhabitants in 1500 when Venice, Florence, and Genoa, its principal competitors, were very much larger [(.Krekic 1972{:\054-55}.)]. Florence probably had over 60,000 inhabitants and Venice contained about 100,000. With a patrician population of about 400 adult males alive at any one time, it was about the size of a community that an anthropologist would study in the field.
Second, the archival sources are incredibly rich. As Braudel has observed, the Ragusan Archives are far and away the most extensive of any city in the Mediterranean. To anyone with the time and patience to study the voluminous Acta Consiliorum, they afford an opportunity to observe the extraordinarily well-preserved spectacle of a medieval town in action [1973: 1258-9].
Not only are these series very extensive for the mid-fifteenth century, but the fact that there are many different series about so many different aspects of political and economic life, makes it possible to study this community `as a whole'. These two features, very rich sources on a small but important community, makes Ragusa an excellent subject for anthropological study1.
Third, the Ragusan material also permit us to witness the operation of a miniature but important centralized state. The wealth of patricians families, and indeed of the city-state as a whole, derived largely from long-distance trade. It is paradoxical that this prosperous city is not particularly well located on the Adriatic coast for sea-borne trade. Ragusa lies on the open sea, beyond the Dalmatian islands which give ready shelter from treacherous Adriatic storms. Most of the trade of the period plied the Dalmatian side of the coast since the Italian coastline is largely devoid of islands, harbours, and the shelter they provide. Ragusa was somewhat better situated with respect to the Balkan interior. But the immediate hinterland could not support a growing population. Indeed, the interior was the domain of warring Slavic nobles who periodically disrupted trade and threatened the city. The Ragusans preferred to cling to the fringe of relatively unproductive land along the Adriatic. The 425 square miles of territory might be suitable for vineyards and salt pans, but they could not provision even a small city of about 6,000 inhabitants. Thus, while the patrtician e\*'lite owned extensive estates, their wealth derived largely from their activities as merchants, not as rentiers.
Members of the patriciate both controlled and staffed the state apparatus. This apparatus not only provided a jural framework for insuring contracts and adjudicating complex commercial disputes but also determined the volume, direction, type, and terms of that trade. The various agencies collected customs duties, negotiated the terms of trade with foreign powers, established trading centres, awarded important trading contracts, and so on. The various functions of government were executed by adult patrician males sitting as a Major Council and, elected from this body, two smaller councils and various state offices. Some of the offices were directly concerned with trade and its regulation, some with civil defense and provisioning, some with the law and courts, some with the stewardship of various religious and charitable bodies, some with the administration of outlying areas, and so on. But the principal decision-making functions were reserved for the Major Council and three other institutions: the Rector, his Minor Council, and the Council of Appeal (or Senate).
In 1332 the Major Council (Consilium Maius) of Ragusa closed its membership. Henceforth only members of this council could hold state office and this action defined the effective limits of the patriciate. The research project has focused on two dimensions of life in this patrician community, the structure of government and ties of kinship and affinity. The primary source of information on government are the records of the various governing bodies. They are very extensive. The Acta Consiliorum referred to by Fernand Braudel contain about 500 manuscript pages for each year. In this study I have concentrated on the period from 1440 to 1460 -- a total of about 10,000 manuscript pages of source material on the three governing councils. The number of Major Council members fluctuated from year to year but it averaged about 400 adult males. The Major Council established governing ordinances, determined policy, and it elected the numerous officials who staffed the state apparatus. In any given year there were at least 152 separate posts to be filled from the ranks of the patriciate.
While a great quantity of information about politics and government have been recorded, the computer has been used primarily to store and analyze information about office-holding and elections to office. This latter type of information is relatively easy to extract from the manuscript sources and `put up' on computer since it appears in standard format in the records. It also lends itself to coding in compact form. All patricians holding office between 1440 and 1460 have been recorded, as well as all elections held in the years 1440, 1445, 1450, 1455, and 1460. This is discussed in more detail below.
Materials on kinship and marriage come from a more diverse set of materials. Much is contained in two sets of marriage documents. The first of these, the Pacta Matrimonialia, is a marriage contract in which the potential spouses promise to have one another as husband and wife. This document contains the names of their fathers and often the names of other defined kinsmen. The Carta Dotalis contains the amount of dowry given to the bride. It, too, has much information about marriage and kin relations2. Further information is contained in the minutes of governing councils, wills, contracts of various types, and so on3.
At the heart of the project are several large data files containing information about kinship and marriage. The first of these is called GEN (after GENealogy)4. It is an eight field database (Table 1)5. It includes given name\**, There is a separate file on NAMES, as well as certain other files, which are not discussed here. The process of identifying particular individuals and distinguishing between individuals in archival sources is too complex a topic to be dealt with in any detail here. Suffice it to report that NAMES contains all patrician names and a personal identification number, where the latter can be determined. There is a separate glossary of names and a command file which checks for equivalents. In preparing the information of computer processing, unknown or problemmatic names are `dumped' for manual checking. For an excellent example of how such a problem can be handled using artificial intelligence techniques, see {(.Bourlet Minel 1986.)}. status (as patrician, illegitimate or otherwise), and date of death of any individual mentioned in the documents. Since dates of birth are unknown and can only be estimated, no special field is set aside for this information.
GEN also includes a field called `Hack' containing genealogical information6. This field of genealogical data, the `Hack number', consists of a series of two digit numbers. The first two digits encode the individual's clan affiliation (Diagram 1)7. The following pairs of digits encode generation level and birth order. Blasius, reading upwards from bottom of the genealogy, is the third child of the first child of the second child of the third child in the Zorzi clan. Thus, Blasius' genealogicial code would be (moving from the top of the genealogy down): 3303020103.
In addition, GEN also includes a unique four digit identification number for each individual. This ID serves two purposes. It allows us to include individuals in the files whose precise genealogical position may not be known. It also gives a unique identifier which can be used to link GEN to other files in a relational database8. There is an additional field in GEN which includes the individuals' mothers' ID. Finally, there is a comment field which contains a pointer to a comment file on the individual. Since these files are terse by design, much textual material and additional information is kept in text files `pointed at' and located by the entry in the comment field. At present GEN contains information on about 3000 individuals.
Information on marriages is contained in two files. One, called PAC (from PACta Matrimonialia), contains skeletal details of the matrimonial contract. This contract formally committed the couple to marriage, specified the time when the marriage would be consummated, indicated whether papal dispensation had been applied for if the marriage transgressed the Church's limits on intermarriage, outlined details of the dowry, and so forth. Like GEN in being a terse and highly structured data file, PAC contains critical information on the matrimonial contract and pointers to files where more verbose and elaborate information is stored (Table 2). PAC also contains the ID numbers of bride and groom. These ID numbers point back to GEN or on to the second marriage file. If either party had guardians or tutors, PAC also contains pointers to files containing their names and other information.
The second marriage file is called CAR (from CARta Dotalis). It records skeletal details of the eventual dowry settlement (Table 3). The often very elaborate incidental information on the various loans, sales, and other arrangements that were required for paying the dowry are once again pointed at by a comment field.
The database on kinship and marriage has been developed by breaking down some rather large and cumbersome flat files. The information contained in these files was abstracted from archival sources, coded onto code sheets, punched onto cards, and then loaded onto computer. This was a very time-consuming and laborious process. The development of the database system has been a further time-consuming process. In future work I would like to use a system which will allow direct data entry of free text. Then, running a database program as a background activity, it would be possible to move through text files placing blocks of text, or data abstracted from the text, directly into the database.
Moving from the files on kinship and marriage to political data, there are two major files. These are both flat files on mainframe (with copies on floppy disk) and they have not as yet been broken down into smaller and more useful records for home processing. The first of these flat files on politics, OFFHOLD, contains information on state offices and their incumbents between 1440 and 1460 (Table 4). Its fields include date of election, the office, name of the newly elected office-holder, the name of the previous incumbent, and (where known) the reason the previous incumbent vacated office. Since approximately 50 offices had to be filled each year, and since many offices had multiple incumbents, this data set is very extensive. It includes over 3500 records.
From government records it has also been possible to record information on elections. Election data includes the names of all candidates nominated for a particular office, the number of votes cast for each candidate, and the outcome of the ballot (Table 5). Since material on elections is particularly voluminous, it was decided to sample this material at five year intervals. For example on 26 February the Major Council met to elect a new Rector for the the month of March. On this occasion four different and inconclusive ballots were held. Nine different candidates were proposed but none was able to command a simple majority. On March 5th three more ballots were held before Ser Junius Dobre Calich was elected. Of the total of seven ballots, elelven candidates were put forward, and seven of them appeared on more than one ballot. This was not uncommon. When four of the justiceships on the Criminal Court became vacant in the same year, 68 separate ballots spread over 30 different council meetings were required. A total of 35 candidates from 20 clans wer put forward. All the various elections held in 1440, 1445, 1450, 1455, and 1460 were collected for analysis. These five sets, called ELEC40, ELEC45, and so forth, record over 4400 ballots.
The remaining data file lists all the members of the Major Council sitting between 1440 and 1460 (called MCOUNCIL). It is used mainly for reference.
The advantage of a database is that it allows data to be easily amended and expanded. For example, we may know that a particular individual is the son of another individual although his birth order position in the sibling group may be unknown. He can still be entered in GEN even though the entry for `Hack' will be (temporarily) left blank. When birth order is determined, it can be entered in GEN without making any alterations to PAC, CAR, or other files.
A further advantage is that the data can be stored in small units which can be combined in many different ways. For example, GEN contains a field with the person's mother's identification number (the field is MUMID). From this number it is possible to locate the mother's GEN record. The father's GEN record is located from the individuals own Hackenberg number using various string functions. Using the example given above, by deleting the last two digits from Blasius's Hackenberg number we have Blasius's father's number. From this we can find his GEN record. Being able to locate father and mother in this fashion makes it possible to find any paternal, maternal, or affinal relative.
The task of linking files and extracting particular types of information is accomplished through a set of purpose-built command files written in dBase II command language. At present these constitute a set of primitive genealogical utilities, but they are written so that they can be combined into more powerful search and analysis tools. Using GSX (Graphics System Extension) I am working out routines for interactive display and manipulation of genealogical data. These are purpose-built for the Ragusan data and my own research interests. They have neither the power nor the elegance of the impressive routines being developed by Ryan [1985]. At present, however, my efforts at graphics have resulted in a few paltry successes and a couple of hopeful failures.
The linking of data fields such as identification numbers (ID) and dates (BEGIN, END) between GEN, PAC, CAR, and other files makes it possible to combine pieces of information from different files for special inquiries. For example, we can reconstruct marital careers, determine the offspring of particular unions, place marriage alliances in the context of other marriages, and so on.
To date only OFFHOLD has been analyzed. Almost all of this work has been done on the mainframes at the University of Manchester Regional Computer Centre (UMRCC). First, a series of indexes have been prepared which list the data by (a) name of office-holder, (b) by office, and (c) by date of election. While they have proved very useful in some small analytical tasks, they were prepared initially to use as reference aids in the Dubrovnik Archives. I will find out just how useful they are when I next return to the archives.
OFFHOLD is being used to chart political careers. First, I analyzed career structure looking for the sequence of offices that the `typical' patrician might hold. It was known how these offices ranked in terms of relative status from other sources, but it has been very useful and illuminating to work out the normative pattern. Simple statistical techniques were used to determine which offices were most likely to preceed, or follow, any particular office. I also determined the number of offices, and the number of years, which were likely to intervene between any two given offices. From this analysis the normative order of office-holding can be plotted and also the likelihood of alternative career patterns.
Moving on from this, individual careers are being examined against the backdrop of career structures. While the work is not very advanced, I am particularly interested in seeing individual careers in the context of particular events in a person's life. The same materials, seen in the context of significant events in political life of this city-state, is a long-term goal of this research. More immediately, I shall be juxtaposing several different individual's political careers in order to work out patterns of alliance and opposition in Ragusan politics. This is where the election materials (ELEC40, etc.) will become critical. By looking at people who opposed one another for office, and by examining the outcome of the balloting, it should be possible to determine the relative strength of various contenders and to see how these changed over time and circumstance.
Finally, OFFHOLD will be used to chart the flow of various individuals through state offices and governing councils. From these data it should be possible to see if certain sets of individuals monopolized various sectors of government and to relate these sets to ties of kinship and marriage. From this outline it will be seen that this is a long-term research project. While the major databases are now in place, the processes of expanding and amending them will continue. As I have indicated, they are being used to address problems in kinship and marriage. They are also being used to study politics in an Early Modern European city-state. From these beginnings I intend to bring together kinship and politics and, in the long term, economics as well.
TABLE 1@Data Fields in GEN Field Names@Contents of Field ID@Identification Number HACK@Hackenberg Number NAM@Christian Name SEX@Gender DEAD@Date of Death (Yr/Month/Day) STAT@Status MUMID@Mother's Identification Number COMMENT@Pointer to Comment Text File TABLE 2@Data Fields in PAC Field Names@Contents of Field GID@Groom's Identification Number BID@Bride's Identification Number CONS@Period Until Consumation DOTE@Dotal Amount VEST@Amount for Vestments CAR@Carta Dotalis Date BEGIN@Date of Contract END@Date at End of Contract REND@Reason for End of Contract COMMENT@Pointer to Comment Text File TABLE 3@Data Fields in CAR Field Names@Contents of Field HID@Husband's Identification Number WID@Wife's Identification Number DOTE@Dotal Amount EXAG@Amount in Gold BEGIN@Date of Contract END@Date at End of Contract REND@Reason for End of Contract COMMENT@Pointer to Comment Text File TABLE 4@Data Fields in OFFHOLD Field Names@Contents of Field REF@Source of Information BEGIN@Date of Election OFFICE@Office Held SUR@Surname of Office-holder NAM@Christian Name of Office-Holder FNAM@Father's Christian Number REPSUR@Surname of Person Replaced REPNAM@Christian Name of Person Replaced REPFNAM@Father's Christian Name of Person Replaced REASON@Reason Previous Incumbent Vacated Office TABLE 5@Data Fields in ELECT50 Field Names@Contents of Field REF@Source of Information BEGIN@Date of Election ACTION@Type of Council Action OFFICE@Office CANDI@Status of Candidate SUR@Surname of Candidate NAM@Christian Name FNAM@Father's Christian Name FFNAM@Father's Father's Christian Name PRO@Votes For CON@Votes Against ABS@Votes Abstaining BALLOT@Number of Members Balloting AGENDA@Number of Item on Agenda