STEP 3: AUTHORSHIP BY VERSE-WORDTAG After we've evaluated all the word blocks and compared them to all the possible authors, the blocks are converted back to verses. It's at this stage that we can more easily match the results to versified greek or hebrew text.
Illustration below:
[1] In this example, we're looking at Mark 3:3 (instead of Luke 3:3!). In the second column are the nth word/tag count in the verse.
[2] These are the actual morph-tags for the greek (Mark). Keep in mind this is done for both the strongs numbers (vocabulary) and the tags (syntax).
[3] This column shows the author (book) that has the closest style to the actual morph-tag (not including the actual author). So, for example, in the first data row, the 1st John style best matched the use of 'CLN' in this word environment.
[4] This column is the error percent for the actual author (Mark). So for the first data row, Mark's style closely matched the actual with a 4.625% error. Now if you're wondering how Mark could have 'any' error with his own writing, remember that (a) there could be other embedded authors in Mark, and (b) the neural network is not trained to memorize ... there's always some error allowed.
[5] On the left-5, is the author list and on the right-5 is a list of error percentages. The author list matches the error list. Note that the author list and associated percents can vary by analysis to see the impact of inclusion.
Now, if you look at column #3 (BestAuth), you'll notice that many of the Pauline books closely match Mark at the syntax-tag level. You shouldn't be surprised and I'll discuss this later. Mark wrote about the same time as Paul. The vocabulary results however show that Mark wasn't a Pauline secretary.
Illustration below:
[1] Verse number.
[2] Best author (book) for the verse, not including the actual author.
[3] Error level for the actual author (Mark)
[4] Again, error lists for the alternative writers
At the total verse level here, the best matches are now a mixture of Pauline books, Acts and gospel portions. There's a story here and I'll discuss it in the following pages.
Next page please ... >>