Claudia Celi e Andrea Toschi
Comunicazione presentata alla “HongKong International Dance Conference” il 27 luglio 1990
The research on XIX century theatrical dance in Italy poses peculiar problems to the dance historian not because of the lack of relevant material, but for the opposite reason: the existence of many different sources of information that cannot be readily compared. Even collections of dance librettos give ambiguous data, because not two of them are catalogued in the same way. So, in order to conduct any kind of research, the scholar has to consult one by one the original librettos, when and if they are available to the public. Since the collections are preserved in many different libraries and archives, it would take years of continuous work for a single scholar to consult and collate the data from all existing sources. On the other hand, if the scholars that are active in the various cities collect data from the librettos in different formats according to the occasional needs of the research they are conducting, it is likely that the data will not be comparable, and that much effort will have to be uselessly duplicated.
The experiment we conducted had the aim of assessing the feasibility of a complete cataloguing of italian dance librettos of the XIX century by the adoption of a standard format for the collection of data with a personal computer. This unified format would make possible the exchange of data between the different research centers, and eventually the creation of a data bank containing all relevant data from all existing librettos.
The first step was the creation of the form for collecting data (see fig. 1). The form was drafted with the help of Dr. Sasportes (editor of “la Danza Italiana”) and was used to collect data from the libraries existing in Roma. Since the aim of the experiment was to test the procedures, and not to conduct a complete search, we decided to catalogue only the librettos of the balli given in Roma between 1820 and 1860 that appear in the catalogue of the Carvalhaes collection.
The research was conducted in four different phases:
- Choice of librettos
- Collection of data
- Data entry
- Data processing
The first phase was quite fast, because of the limited scope of the research. After consulting the catalogue of the Carvalhaes collection we made lists of the librettos that had to be examined (85 in number), and assigned them to the two researchers charged with the manual collecting of data.
During the second phase we met an unexpected problem: about 25% of the librettos we wanted were not available for various reasons, so we had to look for them in other collections. As a result we obtained 63 completed forms.
In order to begin the third phase we had to decide the format of the records in the database. This was the most delicate part of the experiment, because the choice of the record format has effects on all subsequent operations. For example if one chooses to have ten fields for the names of the dancers, it’s very complicated to add an extra field for an eleventh dancer in the middle of the job. If, on the other hand, one starts with 30 fields to stay safe, most records will have empty fields, and take away precious space from the hard disk.
These choices had to me made after careful examination of the completed forms and of the problems posed by many of them. All the members of the research group cooperated in order to clear doubts, but in many cases the competence of the dance historian was decisive in clarifying what was actually meant by the words printed in the libretto. Probably the most puzzling problem is the name of the choreographer. It’s often impossible to determine only from the libretto who is the original creator of the choreography. The expression “Composed by” is infrequent and unreliable, because it is sometimes contradicted by other sources. On the other hand the assessment of the original authorship of a ballet is highly controversial in a field where even today there isn’t a well defined copyright. We found ballets that had the same title but were different choreographic versions of the same plot created independently by different choreographers; at the opposite we found two ballets by the same choreographer that had completely different titles but were identical in the plot and in the functions of characters. We decided eventually to take as choreographer the one indicated by the libretto, independently from the fact that he was the original creator or only the reproducer; when we had indications of possible precedents we recorded this in the comments field.
Another problem was posed by the rank of the dancers. Actual indications are rare; sometimes the rank of “prima ballerina” is indicated only by the difference in typographic font in the printing of the name; in other cases a soloist’s rank can be deduced from the indication that the dancer takes part in a “passo a due”, “a tre”, etc., but without a precise specification about the character of the number; in this case in place of the actual rank we indicated the number in which the dancer takes part.
The final format of the database is described in the appendix. In order to save space it has been designed as four linked .DBF files (see fig. 2). This kind of file can be read by most database management programs, so it is independent from the program that we used to develop the procedures (DBIII from Ashton-Tate): the researchers can exchange data and examine them without worrying if they have a different database management program, as long as it can read .DBF files.
We then proceeded to the phase of entering data in the computer manually with the keyboard. At the end of this phase we realized that there was another problem to tackle before we could process the data: the spelling of the names. Even the least difference in the spelling of a name leads the computer to create different entries in listings; for example we had four different entries for the same dancer corresponding to all the possible combinations of Lodovico-Ludovico Pedoni-Petoni. In this and in other cases we chose the most frequent spelling and normalized all other occurrences.
The final phase was obviously the most rewarding, because we could finally use the power of the computer to do all repetitive jobs in a fast and reliable way: we could create indexes according to all possible research keys, and print listings of ballets that correspond to certain requisites. We could analyze the career of a dancer listing chronologically all the ballets in which his name appeared, and thus document the existence in Roma of a corpo di ballo substantially stable over the years. We could also prepare a synoptic table of the activity of the two principal theaters active in Roma from 1845 to 1855 in a form ready to be printed (see fig. 3).
The experiment was not only useful in order to develop and test the procedures for the management of data from librettos; in the course of a critical study of the same period [nota a piè di pagina – C. Celi: L’arivamento de la gran maravija der ballo – Ballo a Roma 1845-1855 to be published in “la Danza Italiana”] the listings and statistics prepared from the database corroborated the results coming from other historical sources. These other sources made clear that the shows produced in Roma were part of a very fast circulation that could bring a guest star to dance in Roma a ballo originally produced in London only two months before. The next logical step is then to extend the research to the shows produced in other italian towns, in order to document the circulation of dancers, choreographers and balli.
Even if the number of librettos that have been examined is only a fraction of the total number of balli given in Roma during the XIX century, the experiment gave a positive answer regarding the feasibility of the project of a data bank for all Italy. A personal computer of moderate cost could store data from at least 20.000 librettos. The problem of merging data from the various sources could be solved if all the researchers that are active in this field adopted the same standard: data could then be exchanged on floppy disks (containing at least 200 records each) or via modem.
We therefore propose that anybody who is interested send us his opinions on the standard we describe, indicating the adaptations that he thinks fit. We hope to be able to organize a meeting in order to discuss all the proposals and unify them in a standard format that can be widely communicated and possibly adopted by all scholars.
APPENDIX
This is a brief description of the main features of the standard.
Data from the librettos is stored in four .DBF files whose structure is described in next page. The four files are linked by cross-referencing pointers. This arrangement has been chosen in order to avoid useless consumption of disk space by the fields that are variable in number. For example the main file (CATBAL.DBF) does not contain the names of the dancers but only two numbers that are the number of dancers named in the libretto and a pointer to the starting record in the file BALL.DBF that contains the actual names and ranks of the dancers. The same method is used for the names of set designers (contained in SCEN.DBF) and the comment lines (contained in COMMENTI.DBF). The procedure for the storing and retrieval of data becomes more complicated, but the space savings is considerable if one thinks that providing empty fields for 12 dancers, 6 set designers and 18 lines of comments would mean doubling the average space consumption.
Struttura del database C:CATBAL.dbf
Campo Nome campo Tipo campo Dim
1 TITOLO Carattere 80
2 TEATRO Carattere 25
3 CITTA Carattere 15
4 DATA Data 8
5 ANNO Numerico 4
6 ATTI Numerico 1
7 STAGIONE Carattere 15
8 GENERE Carattere 30
9 COREOGRAFO Carattere 20
10 N_COR Carattere 15
11 RIPROD Carattere 20
12 N_RIP Carattere 15
13 LIBRETTIST Carattere 20
14 N_LIB Carattere 15
15 AUTORE1 Carattere 20
16 N_AUT1 Carattere 15
17 AUTORE2 Carattere 20
18 N_AUT2 Carattere 15
19 PR_VIOL Carattere 20
20 N_PR_VIOL Carattere 15
21 INV_COS Carattere 20
22 N_INV_COS Carattere 15
23 SCENOGRAFI Numerico 1
24 IN_SCEN Numerico 4
25 IMPRESARIO Carattere 20
26 N_IMPR Carattere 15
27 C_SARTO Carattere 20
28 N_C_SARTO Carattere 15
29 ATTREZZ Carattere 20
30 N_ATTREZZ Carattere 15
31 MACC Carattere 20
32 N_MACC Carattere 15
33 BUTTAF Carattere 20
34 N_BUTTAF Carattere 15
35 BALLERINI Numerico 2
36 IN_BALL Numerico 4
37 PROGRAMMA Logico 1
38 SOGGETTO Logico 1
39 BALLABILI Logico 1
40 RIPRESA Logico 1
41 ABB_OPERA Logico 1
42 OPERA_ABB Numerico 4
43 COMMENTI Numerico 2
44 IN_COM Numerico 4
45 FONTE_COL Carattere 30
46 RILEVATORE Carattere 2
47 OPERATORE Carattere 2
48 LIBRETTO Logico 1
Totale: 660
Struttura del database C:BALL.dbf
Campo Nome campo Tipo campo Dim
1 COGNOME Carattere 20
2 NOME Carattere 15
3 RANGO Carattere 20
4 NUM Numerico 4
5 ANNO Numerico 4
Totale: 64
Struttura del database C:SCEN.dbf
Campo Nome campo Tipo campo Dim
1 COGNOME Carattere 20
2 NOME Carattere 15
3 NUM Numerico 4
Totale: 40
Struttura del database C:COMMENTI.dbf
Campo Nome campo Tipo campo Dim
1 RIGA Carattere 60 Totale: 61