Author Interview: Estelle B. Freedman on Oral Histories and Sexual Violence

In OHR’s spring 2023 issue,scholar Estelle B. Freedman shares her findings from a macro-level analysis of oral history databases in order to analyze women’s memories and responses to sexual harassment and violence. Freedman’s research shows the potential of applying digital humanities tools for distance reading to the close reading of intimate details of oral histories. Her article,“’Not a Word Was Said Ever Again’: Silence and Speech in Women’s Oral History Accounts of Sexual Harassment” allows readers to contemplate historical changes in women’s responses, both silent and vocal, to sexual violence.

Please describe your project of mining archival oral history interviews and the methodologies you used to analyze them.  

The Stanford Oral History Text Analysis Project (OHTAP) began with a question: can archival oral history interviews provide a fruitful source for the historical study of sexual violence? My previous study of the political history of rape drew largely on public records such as newspapers and legal documents. I wondered what kinds of historical documentation might help explore past historical silences about sexual violence, including reasons why so many women had not spoken publicly, for so long. I was not interested in conducting an oral history project about sexual violence but in exploring existing life histories that tap into memories recorded over the past half century. My questions arose in part because of reactions to contemporary revelations of past assault and harassment that discounted women’s retrospective accounts, as if delayed reporting discredited their memory.   

Given the extent of women’s oral history projects since the 1970s, and the recent digitization of many of them, interview transcripts seemed promising tools for understanding both silence and speech. Sampling some of the searchable online collections produced intriguing accounts, including personal experiences, family histories, or political commentary. Working closely with Dr. Natalie Marine-Street, director of the Stanford Oral History Program (as well as a score of student interns over five years), OHTAP collected around 2,500 digitized interviews from nine repositories across the U.S. We processed the text and metadata, developed keyword searches and a digital search tool (named Winnow), and identified a subcorpus of around 450 interviews with relevant language. We then completed qualitative coding of the text surrounding the keywords (for example, distinguishing sexual assault and sexual harassment, and personal from political narratives). Students develop tools for matching qualitative coding results and interviewee metadata to explore demographic patterns (such as race and education) and change over time. I began close qualitative reading of the almost 900 discrete narratives these methods located within the interviews.  

One of our major goals has been to enable scholars who are not necessarily skilled in digital humanities to conduct digital searches of large databases of oral history interviews on any historical topics. An article that Dr. Marine-Street and I are now completing explains our mixed methods, including suggestions for standard practices that oral historians and archivists could adopt to expedite large scale digital searches.  

With the quantitative data we can ask about change over time in two ways: the year of interviews, and the historical eras in which narrators lived. By either measure, we find no linear expansion of speech over time but rather troughs and peaks. For sexual harassment by year of interview, for example, the proportion of accounts peaked during second wave feminism, after Anita Hill’s 1991 Senate testimony, and again after #MeToo. Mapping change over time by interviewee birth cohorts illustrates the importance of race in our analysis. For example, the proportion of Black women’s interviews that addressed sexual harassment remained fairly stable over the cohorts, but the proportions of white women’s interviews addressing this subject increased for those born between 1900 and 1959, then declined for the youngest cohort (born after 1960). 

How did the language (and silences) that women used to describe sexual misconduct change over time? 

Qualitatively, in reading the sexual harassment accounts for the period before the term emerged in the 1970s, I was struck by both continuity and subtle shifts in the nature of “bad old days” narratives. Describing the 1930s through the 1960s, interviewees often referred to women’s personal responsibility for the classroom or workplace indignities they described. Recalling the period after the mid-1960s – coincident with early second wave feminism and equal rights laws – the narrators used language and referred to strategies that began to hold men more accountable. The first incidents of women seeking institutional redress appear by the time that the feminist anti-violence movement formed in the 1970s. Beyond the scope of this article, interviews recalling the subsequent two decades detail a range of opportunities for reporting and resisting misconduct, including formal legal action and organized protests. The later interviews also highlight the role of women entering the professions as journalists or lawyers, as well as activists, in reshaping awareness of harassment.  

What can readers learn from your findings that add to the field of women’s history and oral history?  

First, we learned that even when women’s voices seemed absent from public discourse, oral histories may contain significant recollections of a topic that has seemed elusive to historians. Digital search techniques allow us to identify the terms they used at the time and then broaden our exploration in sometimes unexpected directions. Second, oral historians and archivists need to adapt the ways they process digital interviews and collect metadata about narrators and interviews to make possible large scale data analysis. Third, both data analysis and close readings confirm the crucial intersection of race and gender in the ways women experience, observe, and remember sexual violence.  

You show that women consistently implemented resistance measures to fight sexual discrimination. What did you learn about such resistance through your analysis?  

The answer is complicated. First, it is important to recognize that the retrospective nature of interviews conducted since the 1970s undoubtedly influenced accounts of the earlier period. Many women contrasted their past resignation in the face of harassment with their current understanding that they should not have had to put up with it. I found it striking, though, that in the decades before the naming of sexual harassment as a form of discrimination, interviewees rarely portrayed themselves as victims, even as they recalled teachers, employers, or co-workers who objectified, demeaned, or discriminated against them through sexual language and actions. Whether they complained about the extent of such behavior or considered it rare, narrators typically assumed that it was their own responsibility to evade harassment. Some reported doing so resignedly, putting up with classroom discomfort or leaving jobs. Other women proudly recalled their individual strategies for fending off unwanted sexual attention and, by the late 1960s, more of the recalled seeking formal redress.  

How do you anticipate language surrounding sexual misconduct will continue to evolve in the future? 

I’ve concentrated on looking back to understand prior usage – asking what terms preceded sexual harassment, or what terms beyond legal language implied non-consensual sexual relations. To build historical vocabularies, we identified relevant words and phrases within the text surrounding our initial search terms and added them to our searches. Some new language (such as hanky-panky) led to further accounts of both consensual and unwanted sexual advances, while other terms (such as ladies’ man) primarily revealed consensual relationships. Legal change clearly influenced how women referred to sexual misconduct, including the term sexual harassment, and law will continue to shape language. In the future, I expect less reliance on a gender binary (including the assumption that perpetrators are male or that “women” share viewpoints and experiences). I would like to think that future oral historians will employ more direct and less euphemistic or coded terminology for the range of human sexual experiences. 

On that last point, what most concerns me about the future is how oral historians will pay attention to subjects once considered off limits, including sexual or other kinds of trauma, for their questions will partly shape the historical record. For the history of sexual violence, we need to think carefully about distinguishing between a legitimate sensitivity to the privacy of interviewees and an unconscious or conscious aversion on the part of interviewers to discussing often troubling sexual (or other) subjects. Following the ethics of sexual consent, oral historians can request a narrator’s permission to approach a topic we once considered too private to discuss, in a way that presumes neither a willingness nor a disinclination to do so. Several of the interviewers in our corpus provide examples of listening attentively and allowing space for even painful sexual memories to surface, or to respect silence.

The digital humanities methods employed in your study of archival oral histories have great potential for other topics. What other topics are ripe for analysis? 

So many come to mind – most obviously, other topics once considered highly personal or publicly unmentionable. As a scholar of sexual history these include abortion, contraception, lesbian and gay relationships, and gender non-conformity.  For the history of the mind and body more broadly, personal memories of depression or of body image could enrich historical studies, as could narratives of life events such as childbirth, illness and disability, courtship and divorce, mourning.

If we could standardize processing of digital oral history collections, scholars would gain an enormous database for studying how people remember both their own experiences and their attitudes towards a range of historical subjects, from sports and recreation to culture and politics. I envision a time when archivists and scholars will regularly run digital searches on multiple collections to determine which ones would enhance a particular research project and benefit from digital humanities analysis.