For many years I've heard about the first "stylometric" study of the Authorship Question conducted by Dr. Thomas Mendenhall in 1901. Irttook an extensive and very laborious hand count of the average word length of many Elizabethan writers. It tabulated words as the project continued. Two research assistants were involved. Marlowe, Jonson, Bacon, Shakespeare and other writers were considered.
It has been widely claimed to have proven that Marlowe and Shakespeare were the SAME writer, because it established that both writers have the same average word length.
This similarity in word length has been supported by the far more extensive and "perfect" computer studies by Spavack and Ule, which I think placed it at 4.2 letters per word. Here's Ule's study for Marlowe:
|Ule's finding of the so called Muller Curves for Marlowe and Shakespeare. Nearly identical.|
In and off its self a similarity in word length doesn't seem particularly significant. Why? Because it is possible that the similarity is only in the average word length. Writers with vastly different preferences in word length could have the same average word length. This would happen because Writer A might have fewer smaller words than Writer B. Or the just the reverse.
|Ule's findings for Marlowe's average word length and average frequency of use. From his Concordance.|
So imagine my surprise when I happened over one of Mendenhall's actual graphs. I reproduce it below. If you don't notice two lines, that's because the "two" writers overlapped so closely only ONE line was generated:
|Mendenhall's Study Scanned from John Michell's Who Wrote Shakespeare? page 229|
What we see above isn't that the two writers simply had the same average word length, which could be caused by vastly different vocabularies and words of various lengths. What we see is they have exactly the same average word length because they share exactly the same length words all up and down the line! It was a perfect match! Marlowe and Shakespeare share the same number of one letter words, two letter words, three letter words, four letter words, five letter words and so on right up to twelve letters Talk about an unexpected overlap! No one else from the test group of Elizabethan writers came close. Bacon was quickly eliminated, since his preference is for much longer words, his average an entire letter higher than Marlowe and Shakespeare.
This was an truly astonishing finding and one that should be more widely known and considered.
Ule's stylometric studies of both canons produced a similar result. The writers also parallel each other as closely as they parallel themselves. And Ule's study considered many more factors than just average word length. Ule studied relative prepositional and connective use, sentence length and other quantifiable factors, as well as vocabulary overlap. Here is what that study looked like when two extra plays were added to Marlowe's canon, plays the computer studies suggest as his:
|More statistics from Ule, who publishes about fifty pages of them in his Concordance of Marlowe.|
It is my opinion, however, that nothing is as conclusive as Mendenhall's early study. The two writers are surely one person. The odds against finding two writers that matched continuously right down the line in average word lengths must be extremely remote. Particularly two writers that already have such a remarkable overlap in subject matter.
Mendenhall's study is, thus scientific proof, that Marlowe and Shakespeare were the same writer. It just doesn't get any closer or better than this.
Peter Farey writes that these pairs also have very close average word lengths.
>Charles Dickens and George Eliot
>Thomas Hardy and Andrew Lang
>Lord Tennyson and Robert Browning
>Baroness Orczy and Virginia Woolf
>Oscar Wilde and Conan Doyle
>Mark Twain and Ben Jonson
>Oliver Goldsmith and Jane Austen
>William Wordsworth and Samuel T. Coleridge
>Mary Shelley and Emily Bronte.
>In every case, the profiles are closer than that of Shakespeare's canon to Marlowe's.
Now my question is are we speaking about average lengths here or the perfect overlap that the Mendenhall study found between Marlowe and Shakespeare? We'll have to wait for Peter's answer.
Return to John Baker's Home Page: