It happens often, as it did this month in Hagerstown, Md., when a person asked me, “What about Elvira? What happened to her?”

Elvira and her disappearance are the most mysterious aspect of the Arthur Jordan story. For me, however, Dr. Gustavus Horner has second place locked down.

As I recounted in Condemned for Love, Horner, a local physician, ran from his home to the Warrenton Cemetery on the morning of Jordan’s lynching. He made a sketch of the suspended body but also wrote about the incident in his diary.

His drawing is an important artifact, and the diary could be too if you could read it. Both are housed at the special collections library at the University of Virginia.

“The pages of the diaries are extremely difficult to read because of the bleed-through of the ink over the years,” wrote one of the librarians there.

The librarian was being too kind. The more important problem is Horner’s terrible handwriting. His is a perfect example of doctors’ illegible script.    

I struggled to translate the three pages, as did Kirk Goolsby, a Warrenton resident who took an interest in the Jordan case. We teased tantalizing bits from the scribble but often little more than isolated words or phrases.

But I have never stopped thinking about the diary and the clues to Jordan’s lynching that it might hold.

This month I asked one of the professors at the University of Mary Washington, an expert in artificial intelligence, if he could help.

He used GPT-4, an AI program designed for “natural language processing and text generation,” according to the manufacturer. The result was gibberish such as this: (The question marks are his)

[?] place the evidence it is to
[?] it is concluded all essentially
[?] of all this [?] [?]

“I am not aware of any publicly available tools that would be able to do this,” he added.

I also asked Gianluca De Fazio, a professor at James Madison University, if his students would like to tackle the puzzle. De Fazio has written about how his students worked with the Library of Virginia to transcribe old lynching records and place them online. He looked at the photo of one of the diary pages, however, and said he didn’t think his stdents would be able to transcribe it.

De Fazio did recommend that I crowdsource the job at From The Page, a “wonderful website,” he said, where volunteers work with libraries to transcribe old documents.

Sara Brumfield, one of the creators of the website, invited me to post the pages, though she was not optimistic. “We are not transcription experts, just software experts who build a transcription platform,” she said.

Even so, one of their volunteers appears to be intrigued. I have received automatic notices from the site that one of the volunteers has opened the pages. No word yet on any results.

Brumfield also recommended Alabama Memory, a project at the University of Alabama that seeks to remember the state’s 800 lynching victims.

Dr. John Biggie, director of the project and an associate professor of history and African-American studies at U.A., invited me to send him the photos of the diary pages. Again, no word on any results.

Stay tuned. Horner’s handwriting may be too tough for a computer program, but I’ll bet there’s a Sherlock out there who can help.