Abstract

“Forced Labor 1939-1945” commemorates the more than twelve million people who were forced to work for Nazi Germany. Nearly 600 former forced laborers from 26 countries tell their life stories in detailed audio and video interviews. The interviews have been made accessible in an online archive at www.zwangsarbeit-archiv.de. Sophisticated retrieval tools enhance a user-friendly research and a close-to-the-source analysis of the interview recordings.

Introduction

National Socialist Germany created one of the largest systems of forced labour in history. Not until 60 years after the war did the debate over compensation help to revive the memory of the long-forgotten victims of forced labour. While the survivors strive for a living remembrance in the country where they lost their youth, historians increasingly ask about individual and collective patterns of memory. Today, however, only a few survivors remain to recount their experiences. To preserve and to present some of these testimonies to a wider audience is the aim of the digital interview archive “Forced Labor 1939 – 1945. Memory and History” (www.zwangsarbeit-archiv.de). It is dedicated to remembering the more than twelve million people that were coerced to work for Nazi Germany.

After finishing the task of financial compensation to former forced labourers, the Foundation “Remembrance, Responsibility and Future” has initiated and financed several remembrance projects, among them a collection of nearly 600 narrative life story interviews with former concentration camp inmates, prisoners of war and ‘civilian’ forced labourers. These testimonies were recorded in 2005 and 2006 by 32 partner institutions in 26 countries under the coordination of Oral History expert Alexander von Plato at the Institute of History and Biography at FernUniversität Hagen.[1] Among the interview projects were academic institutions like the Yale University or the Bulgarian Academy of Sciences as well as experienced project groups from non-governmental initiatives like Memorial Moscow, Berlin History Workshop or Živá pamět’ Prague.

Subsequently, a joint project between the Foundation “Remembrance, Responsibility and Future”, the Freie Universität Berlin and the German Historical Museum was formed in 2008 with the aim to safeguard and provide convenient access to these multilingual audio and video interviews in full length and accompanying materials for research and education purposes.

The Interview Collection

The definition of forced labor underlying the interview collection must be considered in a broad sense of the term, comprising bonded work under German civilian, military, police or SS control. Most of the interviews took place in the Ukraine (82), Poland (82) and Russia (70).[2] About one-third of the interviewees were “slave labourers” in concentration camps; 134 witnesses were Jewish survivors, and 46 Sinti or Roma. A total of 590 witnesses were interviewed, 341 men and 249 women.
The comprehensive life story interviews were recorded as video in 192 of the cases, with 391 of the interviews on audio, and took three to four hours on average. Transcripts and short biographies, and in many cases also photos, documents and further accompanying materials were supplied with the audio and video tapes. All in all, over 4800 scans of personal documents and historical of current photographs supplement the recordings of over 2000 interview hours in total.

Starting from this material, which is preserved in its original form by the German Historical Museum, the Center for Digital Systems (CeDiS), the Competence Center for E-Learning, E-Research and Multimedia at Freie Universität Berlin, began with the creation of a digital archive and educational material on DVD and in print.

Educational Materials

Since direct encounters with the few surviving witnesses are getting ever more difficult to realize even though pupils respond very well to these lively accounts, the digital educational material “Video Testimonies for School Education: Video DVD – Learning Software – Teachers’ Book” has been developed. The DVD centers around the life stories of five former forced laborers and includes their testimonies cut as 25-minute biographical short films with accompanying interactive learning materials, as well as two background films, specially animated maps, timelines and a glossary, documents and photos. The German language material is being distributed by the German Federal Agency for Civic Education (Bundeszentrale für politische Bildung).

Creating a Digital Interview Archive

In order to realise the online archive, the wealth of audiovisual material had to be digitized, the information provided curated and normalised for a project-wide database, the transcripts aligned with the interview media through a process of segmentation, and the interviews documented and indexed for advanced retrieval and navigation possibilities. With the sole exception of the digitization, all of these tasks were performed manually by academically trained staff and, where required and possible, native speakers. The intent was to ensure the availability of the interviews online, uncut and preserved in accordance with the original interview context for a wide German and international audience, with equal importance on accessibility to the general public, users in school and higher education and academic research after registration. With almost all interviews being conducted in languages other than German, it became evident that the transcripts and their translations had to be presented along with the interview recordings.

Digitization

CeDiS digitized the video and audio tapes in a one-to-one process using the DIGAsystem software by David Systems. Qualitative post processing took place without cutting or changing the content. The data formats Broadcast WAVE (for 391 audio interviews) and DV (for 192 video interviews) are commonly used intermediate formats. As reference copies they are intended to provide a basis for generating other up-to-date formats for end-use in the future. For the online platform, the files have been being transferred to an internet-compatible format (Audio: MP3, Video: MPEG4 Codec H264 and FLV onVP6).

Segmentation of Transcripts

Segmentation is the process of aligning the transcript (and its translation) with the interview video or audio media via timecodes. The purpose of a segmented transcript is to allow navigation to a specific passage in the interview – for instance, a spoken sentence found via full-text search over the transcript. For the archive “Forced Labor 1939-1945” a flexible segmentation scheme was chosen that follows closely the rhythm of the spoken word and tries to encapsulate sentences. This detailed scheme leads to segments of variable length, often not more than a couple of seconds in length. Segmentation was performed offline using the Inqscribe transcription and subtitling software, and was the basis for all documentation and indexing performed within the interview timeline.

The segmented transcript files were imported into the database of a web application that combined the original interview metadata from Excel spreadsheets, the documents and photos, and references to the audiovisual media files into a central repository. All this information could be viewed in context and easily edited. Any changes to the data were tracked and versioned in order to record the change history from the original dataset. This central repository later evolved into an editorial system for all further work and documentation of the archive contents.

Documentation and Indexing

Just as with segmentation, the documentation and indexing of 400 selected interviews was performed with the help of over 100 researchers over three years who logged onto the editorial system to work with the interviews, supported by a central team at CeDiS who advised the mostly external researchers, and ensured that indexing was performed in accordance with the guidelines set out for this purpose. The entries provided were later standardized into registers for places and persons, with additional classification data for camp types, different production sites of the same company, and enriched with geocoordinates. From the data originating from indexing, the biographic metadata fields such as the interviewee’s stations of forced labor or places of living after the war were freshly populated.
At the same time, the interview chronology and structure was described using two levels of headings to later form a table of contents for the viewer to browse and navigate. In addition, the short biographies and personal metadata were reviewed.

Editorial System

In order to streamline this process of assigning tasks to competent individuals and to coordinate and manage the effort as a whole, a collaborative web platform was developed on the basis of the central repository for work on the materials, data and transcripts and for documentation and indexing. This editorial system also served as an interface to the project-wide database and central reference for all data on the interviews or work processes.

The online tools for documentation and indexing in the editorial system.

The editorial system also included online tools for documentation and indexing of the interviews, as well as full workflow and quality assurance management for these tasks. The web-based indexing environment allowed playback of the interview in tandem with the segmented transcript (and translation thereof), and allocation of structured information on places, camps, companies and persons to the exact sentences they are mentioned in the interview.

Online Archive

The initial online publication in a Content Management System in early 2009 did not allow for systematic search or update of the data and did not offer the advanced faceted search options envisioned for the archive. A platform separate from the editorial system was devised that would contain a more stable set of data exported from the editorial system at intervals. This archive platform was then developed with ease of use and stability in mind, whereas the tools for documentation and indexing and the editorial system itself were constantly improved and expanded upon. While the archive contents are not translated into English at this point the site is available both in a German and English version (with a Russian version planned for 2013).

The project website features rich background information on the subject of forced labor.

The interview archive is embedded in a website which provides contextual information about the collection, the project, forced labor and compensation. Animated maps and an interactive timeline are used to visualize background information. Short extracts from the interviews give an impression of the archive’s content and tackle specific historical issues, for example different experiences of the liberation of the Auschwitz on Jan 27th.

In light of confidentiality of the interviewees, access to the contents is controlled in a comparable manner to a physical, local archive, which would require a user to visit in person. Potential users must apply for registration and provide not only personal details but also a statement about their motivation for access. The often raised concerns about an adequate balance of confidentiality and access in the context of digital or online archives can be adequately dealt with these provisions.

The Online Archive is primarily a presentational platform and does not correspond with digital archive infrastructures[3], mainly because archival methods predominantly supply detailed data on provenance and formal description of the archivals, whereas our concern with the material was a description of biographical points of interest in the context of forced labor told through witnesses’ life experiences.

Combined Search Principles

Essential to the Online Archive are its search functionalities that combine the classification scheme of experience group, deployment area, internment conditions, interview language (among others) with the flexibility of full-text search. The selected search criteria are applied consecutively in a faceted search, with each choice narrowing down the number of possible interviews of interest and showing the remaining options for further drill-down. For instance, users can select to identify everyone who gave his/her interview in “Polish” and belonged to the group of “politically persecuted”. In this case, 32 interviews from the archive would be immediately found. The option to refine the search for those who labored in the field of “industry” at this point shows 13 results to exist for the combined criteria. These systematic categories communicate to the user the breadth of content they can expect – spread of interview languages, or types of interviewee experiences, etc. – at a glance.

Full-text search offers the possibility for individual, specific search interests. While full-text search will also give matches for interview metadata, such as place of deportation, its full potential comes from leveraging the interview transcripts and translations. Full-text search takes advantage of the segmentation of the interview transcripts, giving results in corresponding interview passages, and allows direct access to that sequence in the interview. This nonlinear access is very fine-grained, typically linking with the exact sentence.

The highlighting of transcript results helps to remedy some of the disadvantages of full-text search, such as little to no indication of the relevance of word matches to the interviewer’s experience. Places like Auschwitz are sometimes mentioned as a symbolic reference or point of comparison without any actual experience of the interviewer. By showing the corresponding transcript passage already as part of the search results, the user already can choose to ignore obvious mentions of little relevance.

Simultaneous presentation of interviews, transcripts and table of contents in the online archive.

Interview Presentation

Visiting an individual interview page, users are presented the option of playback along with all the metadata and table of contents laid out on the screen simultaneously. As the interview is played back, the transcript is shown below the media player as text passages that are more substantial than film subtitles. Progress through the table of contents is shown via highlighting of the currently playing section. The table of contents serves multiple purposes: providing an overview of the life experience and the narrative structure of the interview, as well as enabling direct navigation to the corresponding interview passages. Respecting the original interview’s structure, multiple mentions of the same narration topic are not omitted in the headings but labeled as recurring.

The overview that the table of contents provides is important in offsetting the ease of directly retrieving text passages via full-text search, where otherwise users might only engage with the specific passages cited and then go on to the next result in their search, without due respect to the witnesses effort in narrating their life experience in these testimonies.

Summary

The online archive “Forced Labor 1939-1945” offers a highly user-friendly interface and helpful research tools. The capacity to locate thematically relevant segments, or particular names and terms within the interview has been made possible by an immense input of time and personnel, which has produced, among other things, the complex indexing of the content of every interview, including producing transcripts and translations as well as registers and headings. In the process, an editorial system was developed as a collaborative web environment for indexing of segmented audio and video content, and as a central repository and workflow system for the project. In the archive platform, a search engine was realized to combine faceted search criteria with full-text retrieval. The segmented transcripts and headings are used in tandem with the interview media to create an interactive, non-linear multimedia environment for the witnesses’ testimonies.

Jan Rietema, IT Multimediale Archive, Freie Universität Berlin.

[1] Alexander von Plato/Almut Leh/Christoph Thonfeld (Eds.): Hitler‘s Slaves. Life Stories of Forced Labourers in Nazi-Occupied Europe. New York/Oxford: Berghahn 2010.

[2] Other countries: Belarus 42, Czech Republic 40, USA 30, Israel 28, Slovenia 24, France 21, Romania 17, Hungary 15, Croatia 12, Spain 11, Slovakia 11, Serbia 11, Norway 11, Netherlands 10, Italy 9, Bulgaria 9, South Africa 8, Lithuania 8, Germany 8, Bosnia 8, Moldavia 6, Macedonia 6, England 6, Latvia 5.

[3] By supporting electronic finding aids in EAD (Encoded Archival Description), for instance.