Shakespeare Statistics

Authorship of Thomas of Woodstock or Richard II, part 1

BRD-Flagge auf deutsch               All data were generated from R Stylo (see: Computational Stylistics Group Homepage).

MF1W, MF2C and MF3C

Each chart gives the name of the play, the window size, the step size, the type of the most frequent variables (here words = MFW1-grams, character bigrams (MFC2-grams), and character trigrams = MFC3-grams). The culling value is also mentioned. N = 70 refers to the percentage of variables that have been taken into account. It should be noted that MF3C accounts for more text than MF1W and, due to the larger population, is statistically more sound.
There is a continuous display of growing window sizes from 1000 words up to 5000 or 6000words, starting with MF1W, followed by MF2C
and MF3C. The smallest stylistic differences between the reference texts and Thomas of Woodstock give a clear indication of Samuel Rowley up to
IV,2 and of Shakespeare from IV,3 onwards. Taking into account the controversial assignments of Michael Egan und MacDonald P. Jackson (Link)
one can conclude that rolling delta confirms Samuel Rowley as the author of the play in collaboration with William Shakespeare. This finds support
in the fact that later reference texts were stylistically far away from Thomas of Woodstock, but a later revision of the play by Shakespeare is very likely.

The cross-table below displays the window sizes horizontally and the play in its number of words vertically. Window sizes around 5000 words are
supposed to mirror the collaborative situation. In theory they might grow until they cover the whole text. This would be Burrows's original delta.
The major part of the play is indeed by Rowley, but Shakesepare seems to have finished the play. The table starts off with 500 words (B6) which is the
first measuring point of the 1000-word window. The next measuring point is at 750 words, covering words 251 to 1250 of the 1000-word window.
But 750 words is also the first measuring point of the 1500-word window (C7). As the window sizes grow the first measuring points move one
column to the right and one line down (towards the end one line up). The last column before the acts and scenes reproduces 2-word collocations,
but there are only 187 variables.