29-02-2016, 03:19 AM
(28-02-2016, 08:34 PM)Sam G wrote: You are not allowed to view links. Register or Login to view.[...] I've wondered if the scribe (who I think was a copyist) had trouble distinguishing between <d> and <j>, and if there should actually be more instances of <j> in the text.
If there was a scribe / copyist, and particularly if he had no opportunity to check back with the person who created the original or draft, then all sorts of interesting things could have happened.
Just imagine that Eva-k and Eva-t were one and the same character, but not written very consistently. The scribe could have decided that these were two different ones, and always wrote it clearly as one or the other.
D'Imperio makes a point about this: they do appear to be written clearly as one or the other, yet in the 4 times repeating 17-character sequence there is one character which is written twice with the extra 'hook' and twice without, as if they are equivalent.
I can think of arguments both for and against this idea.....
The most obvious argument 'for' is the near-replaceability of the two characters. The 50% figure that was computed before can be checked statistically. I've been too lazy to do it, but it would work as follows.
Assume that they are the same, and replace all 't' by 'k' in the MS.
Now make the list of all words with a 'k'. (One can decided whether or not to count the embedded gallows as well).
All the words in this list that appear only once in the MS will add to the fraction: 'not replaceable'.
The number of such words can be guessed from Zipf's law, but could also be counted.
Of all words that appear twice, statistically half will be replaceable and half 'not replaceable'.
Of all words that appear three times, statistically one fourth will not be replaceable, etc. etc.
This assumes that the choice k vs t was made by the scribe with a 50/50 probability. With a different ratio, other fractions emerge.
It would seem interesting to see how the ratios behave especially for these lower-frequency words, and how close the ratios are to the expected figures for a random process.