Attributing the Bixby Letter using n-gram tracing

Jack Grieve, Isobelle Clarke, Emily Chiang, Hannah Giddeon, Annina Heini, Andrea Nini, Emily Waibel

Research output: Contribution to journalArticlepeer-review


There is a long-standing debate around the authorship of the Bixby Letter, one of the most famous pieces of correspondence in American history. Despite being signed by President Abraham Lincoln, some historians have claimed that its true author was John Hay, Lincoln’s personal secretary. Analyses of the letter have been inconclusive in part because the text totals only 139 words and is thus far too short to be attributed using standard methods. To test whether Lincoln or Hay wrote this letter, we therefore introduce and apply a new technique for attributing short texts called n-gram tracing. After demonstrating that our method can distinguish between the known writings of Lincoln and Hay with a very high degree of accuracy, we use it to attribute the Bixby Letter, concluding that the text was authored by John Hay – rewriting this one episode in the history of the United States and offering a solution to one of the most persistent problems in authorship attribution.
Original languageEnglish
Pages (from-to)493–512
Number of pages37
JournalDigital Scholarship in the Humanities
Issue number3
Early online date26 Oct 2018
Publication statusPublished - 1 Sept 2019

Bibliographical note

This is a pre-copyedited, author-produced PDF of an article accepted for publication in Digital Scholarship in the Humanities following peer review. The version of record Jack Grieve, Isobelle Clarke, Emily Chiang, Hannah Gideon, Annina Heini, Andrea Nini, Emily Waibel; Attributing the Bixby Letter using n-gram tracing, Digital Scholarship in the Humanities, fqy042, is available online at:


  • : American History, Authorship Attribution, Computational Social Science, Corpus Linguistics, Forensic Linguistics, John Hay, Abraham Lincoln, Stylistics, Stylometry


Dive into the research topics of 'Attributing the Bixby Letter using n-gram tracing'. Together they form a unique fingerprint.

Cite this