Recognising Douglass More and More

With Trump's re-election and the Democratic Party's soul-searching as to why a majority of Americans decided to adopt greater uncertainty, it seems apt to think of historical figures who experienced palpable shifts in the political environments surrounding them - in our case Frederick Douglass (1818-1895), the abolitionist and social campaigner.

Funded by the AHRC, I had the privilege to undergo an international placement at

the Library of Congress's Kluge Scholarship Center earlier this year, where - each and every morning - we'd arrive to a great pile of newspapers reporting precise updates on

Trump's criminal proceedings in New York over hush money payments. His something between beleaguered and seething face became even more familiar.

The project used Automated Text Recognition (ATR), the process of converting images of text into computer-readable format for deeper scholarship, to recognise Douglass's hand and, from there, further study his political thought and personal philosophies.

After British abolitionists, impressed by his tour, purchased Douglass's freedom in 1847 (Murray, 2023), he became an American Civil War recruiter for the Union, as well as the first African-American to be nominated for the Vice Presidency, though largely absent from the process (Philpot, 2020). Douglass did, however, serve as the American minister to Haiti (1889-1891), although his success in this remains disputed (Nwanko, 2005). In any case, Douglass was prolific enough to break into the (now - dear lord) first Trump Administration’s 2017 news cycle, when during African American History Month – Trump, seemingly unaware of Douglas's death in 1895 on the steps of Cedar Hill, Anacostia, described the abolitionist as “an example of somebody who’s done an amazing job and is being recognized more and more" (Reuters, 2017).

Donald Trump makes strange remarks about Douglass from the White House during African American History Month, with then Housing and Urban Development Secretary – Ben Carson. Carlos Barria / Reuters, February 1st, 2017.

Perhaps we were recognising Douglass more, both literally using ATR – at least of his hand, as well as his broader contribution to abolitionist history. Through this work, we showed ATR to be a massive boon for text-based archival research and a means to develop rigorous, responsible, and creative research to further understanding enslavement legacies. In response to Romein et al. (2022), we also provided a method for text extraction of library collections and presented a map for those institutions to move beyond conceptualizing ATR to operationalizing it at the level of collections and research (Padilla et al., 2019: 8).

More specifically, we focused on automatically transcribing Douglass’s 1886-1887 travel diary (hereafter DD, Douglass Diary) due to its atypical nature, with the historian Fought (2017: 255) suggesting Frederick and Helen Douglass’s tour was uncommonly long, making it a rich historical resource for studying 19th century travelogues as well as their encounters with other cultures.

Tesseract OCR was initially used, but struggled with the diary’s layout and variable font, despite integrating machine learning techniques, leading to us making use of Transkribus ATR instead. Even with limited training data (70 pages, 11,500 words), the resulting transcriptions required minimal correction and achieved a character accuracy of 90.60%. Our model was entitled ‘Late Douglass’, shown below within the Transkribus environment - reach out if you’d like access!

Metadata for ‘Late Douglass (1886-1894)’ ATR model within Transkribus

This model and resulting transcriptions led to further text analysis of the collection to consider insights, trends and patterns in Douglass’s language against his published autobiography. Like previous work shown on this blog, this approach enabled the charting of whether more intimate unpublished and handwritten materials cut against or cemented themes seen in autobiographies written for public consumption.

Though we can't go into too much detail just yet (watch this space!), one shareable finding is that, in following Jockers's (2014: 1-10) method for R word count frequency, Douglass's first autobiography The Narrative of the Life of Frederick Douglass (1845) proved closest lexically to his 1886-1887 diary, despite having the greatest date range: with his own published works being from 1855 and 1881. In taking the most prevalent tokens, we identified a shift in theme - with "slave/slaves", "slavery", "master" being prevalent in his published autobiographies but absent in his diary. That said, although his travel diary denotes more leisurely travel than his previous visit to Britain as a fugitive, this work forms a continuation of his personal history, reflexive thought and experience, with a close reading uncovering allusions to his past bondage, especially surrounding Douglass's inability to pin down his own age (DD, 1887, pg. 35). We, therefore, see the need to measure quantitative text against against close historical research.

To conclude a slightly rambling blog post, which could do more to critique text analysis and ATR (a subject for another post perhaps), as well as spotlight the role of the women in American abolitionism (outside of Douglass), I remain very grateful to Kluge Center and find myself wondering how the US cultural heritage sector will now operate. As we question why a lack of morality appears as more of a qualification for the presidency, there is worth in retaining the value of Douglass as author, intellectual and nuanced (often conflicted) moralist. There also remains worth in advocating for accessible collections-based data to further scholarship and challenge false narratives politicians like Trump expound. Tight grip for the next four years.

References -

Fought, L. (2017) Women in the World of Frederick Douglass. Oxford: Oxford University Press.

Jockers, M.L. (2014) Text Analysis with R, For Students of Literature. Cham Switzerland: Springer.

Murray, H.R. (2023) ‘Frederick Douglass in Britain and Ireland’ http://frederickdouglassinbritain.com, accessed

24 May 2024.

Narrative of the Life of Frederick Douglass, An American Slave Written by Himself: Nota Bene Edition (2001), John W. Blassingame, John R. McKivigan, Peter P. Hinks (eds.). New Haven, Connecticut: Yale University Press. https://frederickdouglasspapersproject.com/page/narrative2001, accessed 14 January 2024.

Nwanko, I.K. (2005), Black Cosmopolitanism, Racial Consciousness and Transnational Identity in the Nineteenth-Century Americas. Philadelphia, Pennsylvania: University of Pennsylvania.

Padilla, T., Allen, L., Frost, H., Potvin, S., Russey Roke, E., Varner, S. (2019) Final Report - Always Already Computational: Collections as Data. doi: https://zenodo.org/record/3152935#.X6WOf-LPzIU

Philpot, T. (2020) ‘Frederick Douglass’, Oxford Dictionary of National Biography. doi:

10.1093/odnb/9780198614128.013.107532

Romein, C.A., Hodel, T., Gordijn, F., Zundert, J.J. van, Chagué, A., Van Lange, M., Jensen, H.S., Stauder, A., Purcell, J., Terras, M., Van den Heuvel, P.,, Keijzer, C., Rabus, A., Sitaram, C., Bhatia, A., Depuydt, K., Afolabi-Adeolu, M.A., Anikina, A., Bastianello, E., Benzinger, L.V., Bosse, A., Brown, D., Charlton, A., Dannevig, A.N., Gelder, K. van, Go, S.C.P.J., Goh, M.J.C., Gstrein, S., Hasan, S., Heide, S. Von der, Hindermann, M., Huff, D., Huysman, I., Idris, A., Keijzer, L., Kemper, S., Koenders, S., Kuijpers, E., Rønsig Larsen, L., Lepa, S., Link, T.O., Nispen, A. van, Nockels, J., Noort, L.M. van, Oosterhuis, J.J., Popken, V., Estrella Puertollano, M., Puusaag, J.J., Sheta, A., Stoop, L., Strutzenbladh, E., Sijs, N. Van der, Spek, J.P. Van der, Trouw, B.B., Van Synghel, G., Vučković, V., Wilbrink, H., Weiss, S., Wrisley, D.J., and Zweistra, R. (2022) ‘Exploring Data Provenance in Handwritten Text Recognition Infrastructure: Sharing and Reusing Ground Truth Data, Referencing Models, and Acknowledging Contributions. Starting the Conversation on How We Could Get It Done.’ doi: 10.5281/zenodo.7267245