Presner Weighs Ethics of Digital Holocaust Archives

By Phyllis Jeffrey - What ethical issues arise when the Holocaust is digitized—when experience is turned into interface, database, and algorithm? Professor Todd Presner, of the University of California Los Angeles, explored such questions on September 29, 2016, in a lecture entitled “The Ethics of the Algorithm: Probing the Shoah Foundation's Digital Archives of Holocaust and Genocide Testimony.”

The drive to avoid what Primo Levi called the nightmare of “the unlistened-to story” has motivated focus in Holocaust and trauma studies on testimony and witnessing. The impetus to honor by witnessing has also played a key role in projects to record and preserve survivor testimony, including the over 53,000 testimonials recorded, archived, and indexed by the Shoah Foundation at the University of Southern California.

The Shoah Foundation archive gestures toward another task of Holocaust memorialization: that of expressing and honoring scale. To “witness” the more than 105,000 hours of recorded testimony would require listening for 12 hours a day for almost 24 years. As Presner explained, it is this unique capability of the digital to communicate the near-incomprehensibility of the scale of Nazi genocide that has led not only to creation of archives in digital form, but also of digitized memorials to victims--the Joods Monument, for instance, a memorial to Dutch Jewish victims of the Holocaust.

But between honoring individual testimony and preserving the scale of tragedy, do potential dilemmas lurk? Presner proposed that we need to look more closely at how an archive like the Shoah Foundation’s goes from the “back end” of the collection (the unseen criteria used for indexing and coding the archive) to the “front end” (a single recording of individual testimony that a researcher can locate, access, and listen to). How, and by whom, are organizational and coding decisions made? What are the ethical consequences, both for honoring individual testimony and for informing the historic account?

Hidden decisions

For Presner, the indexing of survivor testimonials is potentially problematic because the very process used to render the database searchable gives a false appearance of finality, or objectivity. In social science research, he observed, we usually analyze data in order to organize it into an intelligible narrative. The peculiar thing about the Shoah Foundation archives is that they do the opposite—they compile narratives into data.

This process involves a language of categories and sub-categories, based on common themes (e.g. “captivity”), and the coding of each testimonial in one-minute increments. The problem, Presner said, is that although the database has a veneer of neutrality, it contains hidden decision-making moments: the initial selection of “parent” categories and “child” sub-categories; the decision that testimonies ought to be coded at 1-minute intervals; the process of applying category codes.

One possible outcome of the indexing process is that survivor accounts appear more alike than they really are. Analyzing the testimonials of Holocaust survivors, Presner found that something like a “normative” narrative exists. Certain category labels are more likely to be attached at particular minute marks in testimony, and particular categories are more prevalent than others. While this might simply represent genuine commonalities, it may also be the case that common experiences are amplified, not only by the “standardized” approach of original interviewers, but also by patterns in subsequent coding decisions. In other words, the very existence of apparently quantifiable, objective data patterns could lead to the normalization of a certain kind of account—one that becomes privileged over less common narratives.

Concentrating on outliers

More optimistically, Presner also suggested that computational analysis might enable us to see dimensions of survivor experience that we otherwise might miss. To illustrate how the digital could be used to discover rather than normalize, Presner offered his own attempt at a more democratized use of the Shoah Foundation archives. This involved presenting a network visualization in which testimonials, appearing as nodes, were shown linked to others by shared category terms. Presner concentrated on the outliers: those with few nodal connections, visible at the edges of network, whose stories had less common aspects.

He pointed to a testimonial by someone whom the coding indicated came from a less-common geographic location, and who attended Christian religious services, indicating perhaps a new issue for research: experiences or consequences of religious conversion.

By seeking out the less-common narratives to complement stories with more frequent themes, one can, Presner proposed, begin to combat the forces of normalization. Work toward ensuring that “aberrant” narratives are heard is also ethical work, because it democratizes our attention to survivor experience, and ensures that we do not visit on anyone Levi’s nightmare of indifference.  

This event was hosted by Jewish Studies and co- sponsored by the Davis Humanities Institute, Cultural Studies, the German Department, and the Human Rights Program.

Professor Presner’s talk was based on a chapter from the forthcoming book Probing the Ethics of Holocaust Culture.

Learn more about Todd Presner at his webpage.