 The
Library of Congress >> Especially
for Researchers >> Research
Centers |
|
 |
|
|
|
| Home >> Comintern >> About Database | |
|
|
Communist International (Comintern) Archives Project
The Comintern Archives Database: Bringing the Archives to ScholarsRonald D. Bachmanformer Polish Area Specialist, European Division As one of nine institutions participating in the International Committee for the Computerization of the Comintern (INCOMKA), the Library of Congress (LC) played a primary role in making the archives of the Communist International available to researchers around the world. LC assumed responsibility for converting some 175,000 personal names from Russian Cyrillic to their standard spelling in American English usage and translating nearly 20,000 keywords from Russian to English. As a result, researchers who do not know the Russian language have access to the vast Comintern database, including more than a million pages of digitized manuscripts. The Communist International (Comintern) was established in March 1919 to foment world revolution. Within a few years, communist parties existed in nearly all the countries of Europe and by 1930 in most countries of the world. These generally small, often illegal parties looked to Comintern headquarters in Moscow for support and guidance. After the dissolution of the Comintern in 1943, the Communist Party of the Soviet Union (CPSU) took custody of the organization's records. The Comintern archives constitute an important resource for the study of world history during the inter-war period and early years of World War II. The archives also contain interesting biographic material from the Cold War era, when the International Department of the CPSU continued to add personal files. The formerly secret Comintern archives were opened to the public in late 1991. Until recently, access to the collections required a trip to the Russian State Archives for Social and Political History (RGASPI) in Moscow, where researchers faced an endurance test to locate specific information among more than 20 million pages of documents. Extensive finding aids to the collections existed, but they were in the Russian language only. Researchers, regardless of language capability, required staff assistance to determine whether personal files on given individuals even existed. On June 6, 1996, after three years of discussions, the Council on Archives and the Federal Archival Service of Russia (Rosarkhiv) signed an agreement that would make the Comintern archives more accessible to researchers around the world. The agreement set up the International Committee for the Computerization of the Comintern (INCOMKA). Partners in this effort included Rosarkhiv, RGASPI, the Archives of France, the Federal Archives of Germany, the State Archives of Italy, the National Archives of Sweden, the Federal Archives of Switzerland, the Ministry of Education, Culture and Sport of Spain, the Open Society Archives of Hungary, and the Library of Congress. ROLE OF THE LIBRARY OF CONGRESS IN INCOMKAIn the winter of 2004, the European Division of the Library of Congress loaded onto a dedicated terminal in its reading room more than 580 CDs containing the entire catalog of the Comintern archives and more than a million pages of digitized documents. This was the culmination of a multinational effort stretching over several years. In addition to a significant financial contribution as a partner institution, the Library of Congress invested many staff hours to help bring the INCOMKA project to fruition. John Van Oudenaren, chief of the European Division, attended several INCOMKA planning and coordination conferences in Europe and hosted a two-day meeting on linguistic issues in February of 2001. John E. Haynes of the Manuscripts Division also attended meetings, provided expertise in the selection of materials for digitization, served as liaison with 154 historians in 54 countries, and reviewed the lists of Romanized personal names and English translations of keywords. The Library of Congress assumed responsibility for converting some 175,000 personal names from Russian Cyrillic to the Latin alphabet and translating from Russian to English close to 20,000 "descriptors" (keywords/subject headings) taken from Comintern archival finding aids. I was given the task of coordinating this linguistic effort. The goal was to make the INCOMKA database accessible to researchers with little or no Russian capability. The Library of Congress was well suited for this undertaking because of its vast collections of historical, biographical, and lexicographic works and the foreign-language diversity of its staff. Thirty-two staff members, most from the Area Studies Divisions, participated in the project. In the summer of 2000, RGASPI sent to the Library of Congress a list of about 110,000 names taken from personal files (lichnye dela) maintained by the Comintern and, after 1943, the International Department of the CPSU. All names were in Russian Cyrillic as recorded over a period of many decades by clerks with highly divergent levels of foreign-language competence. Our task was to convert the Cyrillic version of the names to their "standard" American-English spelling. The Comintern had files on persons from essentially all the countries of the world as it existed during the inter-war years, including some, like Tannu Tuva, that no longer exist. Many of the persons listed, such as Palmiro Togliatti, were prominent party members, while others were staunch anti-communists, e.g., Harry Truman. There were files on writers, painters, actors, civic leaders, and religious leaders. There even was a file on Karol Wojtyla, Pope John Paul II. But a large share of the persons were unknown functionaries, whose names could not be attested in published sources. The Library of Congress did not have access to the files themselves, which might have provided a Latin spelling for some of the names. The name-conversion process involved four stages. First, using a computer macro devised by our European Division colleague Michael Neubert, we produced a phonetically based transliteration from Russian according to Library of Congress Romanization rules. We arranged the names into more than 100 country tables and distributed them to Library of Congress staff with native or near-native competence in given languages. The German, French, and Swiss INCOMKA partners handled the name conversions for their respective countries. Because of the special problems posed by the Chinese names, Dr. Haynes hired the services of historians at the State Archives Administration of China to identify persons and provide the standard Pinyin transliteration of their names. The Chinese experts, however, were unable to recognize close to half of the entries. In the second stage, Library of Congress linguists analyzed the computer transliterations and converted sequences of letters into meaningful, language-specific combinations. For example, from the Mexican list, KHUAN became Juan; from the Polish list, IATSEK became Jacek; from the Moroccan list, KHADZH became Haj. Library staff attempted to identify individuals and provide the standard American-English spelling of their names, e.g., DZHON RID was identified as John Reed, as opposed to John Read or John Reid. Identifying individuals turned out to be an especially daunting task in the case of the tens of thousands of names originally written in neither Cyrillic nor Latin alphabets. To proceed from the Library of Congress phonetic transcription of a Russian phonetic transcription of a name originally recorded in a third writing system and arrive at the "correct" spelling in American usage was a challenge we were not always able to meet. In the third stage, John Haynes sent the lists to foremost authorities in the histories of the respective countries for "vetting." In most cases, two or three specialists reviewed each list. The lists were in the form of multi-column tables presenting the original Cyrillic, the computer transliterations, and the spellings produced by the Library staff. The specialists confirmed or corrected the spellings, sometimes adding aliases, and returned the lists to the Library of Congress. The final step was incorporating the experts' inputs and delivering the finished tables to the Spanish software company El Corte Inglés, which loaded the information into the special version of ArchiDOC it developed for the INCOMKA database. Just as we were nearing the completion of the name-conversion project, we received a revised list from RGASPI, which created a whole new set of complications. The revised list included tens of thousands of additional names extracted from collective personal files and from the detailed finding aids compiled by RGASPI archivists. Most unfortunately, the revised list merged the additional entries with the original set of 110,000 names. El Corte Inglés, after considerable effort, eventually sorted out most of the new names for us. We completed the name conversion in house, but time did not allow us to send the additional names to outside experts for correction. Meanwhile, RGASPI sent us several long lists of Russian "descriptors" taken from the Comintern archival opisi. Harold Leich, Russian area specialist in the European Division, and I translated these terms into English, and John Haynes edited them. CONTENTS OF THE DATABASEThe INCOMKA project digitized more than a million document pages. Although this is not a small amount of material, it is only a fraction of the Comintern archive of more than 20 million pages. Perhaps more importantly, the project made available to off-site researchers the entire catalog of Comintern archival collections. Even when researchers find that documents of high interest are not among those digitized, the database provides detailed bibliographic descriptions of every delo (file) in the archive, which will enable them to judge whether it is worth traveling to RGASPI or having someone photocopy the material for them. Below is a list of the Fondy (collections, abbreviated F.) and opisi (inventories/finding aids, abbreviated op.) scanned during the INCOMKA project. It is immediately apparent that the focus was Fond 495, the files of the Executive Committee of the Communist International (ECCI). Fond 495 accounts for 57 of the 86 opisi and 7,049 (just over 75%) of the 9,366 files that were scanned. Nearly 85% of the files in these 86 opisi were scanned. The figures in parentheses indicate the number of files scanned and the number of digital images, respectively.
SEARCHING THE DATABASEThe Comintern database delivered to INCOMKA partners in winter 2004 uses ArchiDOC 2.3.7.17 Unicode software. Although searching is rather cumbersome, the software accomplishes its primary purpose -- enabling scholars to search the large database in either Cyrillic or Latin alphabet, identify files of interest, retrieve bibliographic information, and in some cases, view digital images of actual documents. The software provides excellent image-enhancement tools, and readers often will find the digitized documents more legible than the originals. Researchers can print pages of manuscripts for their own use, and the software inserts the bibliographic citation at the bottom of each page -- a very useful feature. The database homepage presents four menu choices: CLASSIFICATION, DESCRIPTORS [key words], PHYSICAL FOND [bibliographic citations, listed in ascending numerical order], and LANGUAGES. CLASSIFICATION organizes the vast Comintern archive into 11 thematic sections, shown below. All researchers who can read Russian, especially first-time users of this resource, will profit from a quick look at the 11 sections. It would enhance database accessibility if the sections and subsections were presented in English as well as Russian, but to make searching via CLASSIFICATION truly bilingual, at least the titles of the 521 inventories (opisi) should be translated. Clicking on the plus symbol before each CLASSIFICATION heading opens a list of subheadings, which, in turn, open sub-subheadings, then opisi, then specific files. Wherever a camera icon appears, a double click will bring up the digital image of an actual document. The CLASSIFICATION sections in English are:
Section 1 might serve to illustrate the logical structure of CLASSIFICATION. Under the heading "Comintern Congresses and Plenary Sessions of the Comintern Executive Committee" there are three subheadings. The first, entitled "Congresses of the Comintern," lists seven Fondy, namely:
Each of the Fondy lists two or more opisi, which in turn list specific dela. For example, Fond 488, opis' 1 lists 18 dela, the first being "Address on Convening the First Congress of the Comintern, 24 January 1919." It happens that this 20-page file was among those digitized, and the researcher can view the document on the screen or print it off. Fond 488, opis' 2 does not subdivide into dela; it is a collection of 76 photographs, which were not digitized. All together, the subheading "Congresses of the Comintern" comprises 15 opisi, and 2,618 dela. Subsection 2, entitled "Plenums of the Executive Committee of the Communist International," comprises 14 widely scattered opisi of Fond 495 and totals 3,130 dela. Subsection 3, "International Control Commission of the Comintern, ICC," contains two opisi totaling 216 dela. Each delo contains, on average, several dozen pages and occasionally can be hundreds of pages in length. CLASSIFICATION provides the most direct path to bibliographic records: a double left click on a title at any level in the CLASSIFICATON hierarchy brings up the bibliographic record in the opposite window of the split-screen display. Bibliographic information becomes more detailed as one proceeds from the section to the delo level. On the section and subsection levels, the record provides a title, the name of the RGASPI archivist who processed the material, and a brief contents note, which generally duplicates the title but occasionally provides a little more information. For example, Section 9, entitled "Interbrigades of the Spanish Republican Army," has the contents note "International formations and brigades of the Spanish Republican Army." On the opis' level, the record additionally provides an information start date and shows the number of dela contained therein. On the delo level, the record contains a contents note, (typically a short paragraph in length) and information start and end dates; indicates the number of pages and the languages of the documents; and provides a list of "descriptors." The list of descriptors can be several pages in length. Comintern database users with poor or no Russian capability are at a severe disadvantage. Although the bibliographic field names can be displayed in English, the contents note is in Russian only. And since it is the contents note (not the actual text of the digitized documents) that is searched by the so-called "text-search" function, persons not knowing Russian do not have access to this useful tool. The researcher without Russian capability who does not have a specific citation to access directly through the third menu option, PHYSICAL FOND, has only one way to identify and retrieve files, the second menu option, DESCRIPTORS. Based on a review of Comintern archival documents, RGASPI staff identified essential terms and grouped them in ten categories of "descriptors." The lists of descriptors can be viewed in either Russian or English. Several of the categories are imprecisely delineated and often overlap. The SUBJECTS list is particularly fuzzy, and many terms are so generic it seems doubtful that a researcher would ever think of searching for them, e.g., Domestic and international situation. Terms within a list are not arranged in any hierarchy of specificity, i.e., there are neither general headings nor increasingly specific subheadings. The distinction between the categories of SOCIAL LEVEL and STATUS is especially vague. The apparent difference seems to be that terms in SOCIAL LEVEL are in the plural form, e.g., graduate students, and entries in STATUS are either singular, e.g., architect, or corporate bodies, e.g., Bulgarian delegation. The descriptor category labels in Russian are not very "descriptive," and their English equivalents (supplied by RGASPI) are even more mysterious. The fuzziness of categories is not a minor inconvenience. Since the software cannot search all descriptor lists simultaneously, the researcher must explore each one separately to have an acceptable level of confidence in the search results. The only apparent advantage of breaking the descriptors up into separate thematic lists is browsability. The inconvenience of having to search several lists should be addressed in future versions of the Comintern database. Descriptors are listed alphabetically, which facilitates browsing. To move rapidly down the long lists (with more than 175,000 entries, the personal names list is the longest), the researcher highlights any descriptor and begins typing a word or phrase -- as much or as little as desired -- and hits Enter. Within an instant, the desired descriptor (or the space where it should appear in the alphabetized list) appears. This "hot search" function is a great time-saver, but it has one major limitation: it is left-anchored. The software does not offer a simple "find in document" function, which would locate any term regardless of its position within a descriptor. The ten descriptor categories (Russian equivalents are in parentheses) are:
Having located and highlighted the desired descriptor, the researcher right clicks to find the option "Show related documents." Left clicking on this option will bring up a list of all dela that include the descriptor in their bibliographic records. Double left clicking on a delo title will display the complete bibliographic record. If the delo was digitized, a camera icon appears before the title. Double clicking on the icon will bring up the document image. If "Show related documents" displays too many (or too few) dela, the researcher can search two or more descriptors simultaneously. This is accomplished by selecting the Search Assistant option on the toolbar, which displays a search form on the opposite half of the split screen. The researcher left clicks on the descriptors (one at a time), and drags them across the screen into the bottom window of the search form. Selecting the option "All of them" activates the Boolean AND operator, while "Some of them" engages the OR operator. The search form also allows one to specify a range of dates or a specific date. The third option on the database menu is labeled PHYSICAL FOND. For researchers who already have specific archival citations, this is the quickest path to the files. Citations are arranged in ascending numerical order, beginning with the Fond number. As with other sections of the menu, one expands or collapses lists by clicking on the plus sign at the beginning of each entry. Part four of the database menu is LANGUAGE. This alphabetic list of all the document languages functions in the same way as a descriptor for searching purposes. A researcher who can read only Norwegian, for example, would click on "Norwegian" and drag the term over to the search window to limit search results to documents written in that language. Researchers who can work with Russian have the option of keying terms directly into a search window and, by toggling to the Latin keyboard, using truncation (the percentage sign) and Boolean operators AND, OR, NOT. ArchiDOC misleadingly calls this function text searching. What the software is searching are file titles and/or file abstracts only a few lines in length. Considering that files often are hundreds of pages long and consist of dozens of documents, the odds are high that a search will show no hits. And even when a search is successful, i.e., one or more files are identified, the researcher's work has just begun. He/she still must page through the file to identify the manuscript(s) where the search terms occur. CONCLUSIONThe INCOMKA project, one of the most ambitious international archival digitization efforts to date, has achieved many, but not all, of its goals. In the near future the database will be available free of charge to researchers through the World-Wide Web, although access to actual digital images will require a paid subscription. We hope the online version will be more user-friendly than the version delivered to INCOMKA partners on CDs. A "search this site" window that would allow users to search simultaneously all ten categories of "descriptors" and to find individual terms regardless of position within a "descriptor" is a badly needed improvement. The conversion of about 175,000 personal names to Latin alphabet and the translation of almost 20,000 descriptors into English were a major undertaking. On the whole, we are pleased with the results. Nevertheless, researchers are certain to discover errors, and we hope the online version will enable them to send corrections and suggestions to the database administrators for updating. Researchers also will soon discover that none of the personal files have been, and probably never will be, digitized -- ostensibly for privacy considerations. But, thanks to the INCOMKA project, researchers now can go to RGASPI with specific citations in hand and request the personal files. The INCOMKA project has made the Comintern archives more accessible, to be sure. But researchers who cannot read Russian still face special challenges and have fewer search strategies than those who can work with Russian. The CLASSIFICATION menu option is available only to Russian researchers. In an ideal world, the titles of the 521 opisi and 230,000 files would be translated into English. At the Library of Congress, we have observed that nearly half of the Comintern database users have had limited or no Russian capability. It bears emphasizing that a large share, perhaps more than half, of the actual documents in the Comintern archives are in languages other than Russian, and that German, French, Spanish, and English account for most of these. These shortcomings notwithstanding, the research community should welcome the Comintern database enthusiastically. |
 Top
of Page |
|||
| |
|
||
| Home >> Comintern >> About Database | |||
|
|
||
|
|||