Anna Bonazzi – PhD Student, UCLA, Germanic Department – annabonazzi@ucla.edu
A Close Look at Holocaust Testimonies: Ngram-based Content Indexing of Survivor Interviews
(Short Paper Abstract)
Holocaust interview collections, such as the USC Shoah Foundation Archive, the Fortunoff Archive, and the Boder archive, are key resources to preserve and share the memories of genocide survivors in the long term. As the size of such collections makes manual analysis impossible, Digital Humanities methods can help categorize and understand their content and narratives. This paper looks at the question of content indexing and argues for the need to develop alternative indexing systems that are not limited to the identification of keywords to summarize content. In particular, the paper proposes the use of a semiautomated DH approach based on Ngrams, or in other words the computation of frequent word groups and sentence chunks. This approach allows us to spot the use of narrative structures that go unseen in traditional keyword-based indexing and reveal patterns in the survivors’ expression of memory and trauma. It also helps shift our perspective from the narrative we expect survivors to use to the actual narrative they use which may go unnoticed if we are not attuned to it.
This paper takes a computational text analysis approach to early Holocaust interviews collected by Prof. David Boder in 1946 and focuses on three elements that become apparent through an Ngram-based indexing method: uncertainty, unexpected topics, and time subjectivity. As for the first point, the use of Ngram helps us see how survivors’ interviews often contain expressions that indicate uncertainty, doubt, reticence, ignorance of events, or explicit refusal to share. Tracking the presence of such expressions in interviews taken over the years and under different conditions allows us to analyze the change of narrative, focus, and re-elaboration of traumatic memories over time.
As for the second point, unexpected topics, I analyze a single interview as a case study to show that a keyword-based indexing approach influenced by specific research questions might miss topics that are in fact quite important to the survivor and come to light more easily with a frequency-based / Ngram-based approach (in this case, the morally ambiguous position of some members of the Jewish community).
For future research, the paper also looks at the way survivors talk about time. While traditional indexing methods group all references to time under the same label, this paper finds that using Ngram-based indexing allows us to bring to light very different types of memory operations if we distinguish between objective and subjective time references (such as “it was 1944” and “we left in the middle of the night”).
In conclusion, this paper aims to show how alternative indexing systems, such as Ngram-based indexing, can help us better understand how survivors select and zoom in on their memories. This method can allow us to better comprehend and communicate how an event such as the Holocaust has to be grasped through multiple, diverging, and nonetheless fully valid narratives, and it can help avoid a circular type of analysis where we only identify the topics and narratives we already expected to find. More in general, it can help us question the standards of a single Holocaust-influenced “genocide testimony” genre towards a more nuanced understanding of multifaceted survivor experiences.

A Close Look at Holocaust Testimonies: Ngram-based Content Indexing of Survivor Interviews

The international conference #DHJewish - Jewish Studies in the Digital Age will bring together more than 60 scholars and heritage practitioners to discuss how the digital turn affects the field of Jewish Studies. The conference is organised by the [Luxembourg Centre for Contemporary and Digital History (C²DH).](https://www.c2dh.uni.lu/) 

[For more information, and a PDF download of the program, please see also here.](https://www.c2dh.uni.lu/events/dhjewish-jewish-studies-digital-age) 



In order to meet other participants and for informal chats in the breaks, join us in our [wonder.me room](https://www.wonder.me/r?id=73a12756-8d93-4871-9374-143253bfbf59).