News

When Professors Speak Out, Some Students Stay Quiet. Can Harvard Keep Everyone Talking?

News

Allston Residents, Elected Officials Ask for More Benefits from Harvard’s 10-Year Plan

News

Nobel Laureate Claudia Goldin Warns of Federal Data Misuse at IOP Forum

News

Woman Rescued from Freezing Charles River, Transported to Hospital with Serious Injuries

News

Harvard Researchers Develop New Technology to Map Neural Connections

Mosteller Joins Federalist Query

NO WRITER ATTRIBUTED

Twelve of the original 85 Federalist papers have never been definitely attributed to either of the principal authors, James Madison, or Alexander Hamilton. Statistical evidence shows that most of the disputed papers were written by James Madison, Charles F. Mosteller, professor of Mathematical Statistics, reported yesterday.

Mosteller is working on a project aimed at eliminating the long-standing problem of whether Hamilton or Madison wrote the unidentified papers by collecting and analyzing statistical data about them.

The major "index of authorship" used in Mosteller's study is a word count. "For instance," he explained, "Hamilton uses the word 'enough' often, Madison only rarely." Mosteller thus might be able to determine that an article with enough "enough's" was probably written by Hamilton, always checking by the use of the recurrent words.

"Only certain words prove helpful," Mosteller added, "and it is necessary to have a substantial theory about variation in word usage. What we call 'contextual' words, like 'war,' 'army,' and 'executive,' cannot be used for discriminatory purposes."

Complicated Style

Another "index of authorship," average sentence length, was employed in a statistical study done elsewhere to determine whether or not a single person wrote the lliad. Mosteller tried this but found that it would not work with the Federalist problem, because Madison and Hamilton wrote sentences similar in length, moreover, they both shared "a complicated style and phrasing."

Mosteller said that his study was probably more interesting as a statistical than as a historical problem since the authorship of the 11 papers is not of particular concern to most historians.

Want to keep up with breaking news? Subscribe to our email newsletter.

Tags