19-09-2019, 10:03 PM
The trouble is that the further you delve into the text, the fewer "real" words you get.
IE, in one paragraph selected at random:
pan ysar choldam yainam cheedy dar ar ainar pody chedy
ainal osos yain qodoaiin ar osor qotal cheeedy sol so
qosoar syainam ytol ol air osy dal al arody alydy dyar
ar kan aimar dol qosoain osar alaim
(bold never appears in the corpus)
tchar kis ckheed pan ar alolol ainol char oshd shar oshd
ochar ar yshd charoshd char dan chol oshdar ain shol shs
shal dain ain octhd char oshd daiin cholchol char oshd
shaldain yshal ochd chy ochar yshd dain cholshol chal an
shaldair yeear chol shol chardan qoshd yshdchar okain an
sholdain char am
From which we see 2 things - longer words in the TT generation tend to looser, and the algorithm used tends to pick up on words and run with them in a way the original scribe didn't (ie shaldain in the second paragraph)
IE, in one paragraph selected at random:
pan ysar choldam yainam cheedy dar ar ainar pody chedy
ainal osos yain qodoaiin ar osor qotal cheeedy sol so
qosoar syainam ytol ol air osy dal al arody alydy dyar
ar kan aimar dol qosoain osar alaim
(bold never appears in the corpus)
tchar kis ckheed pan ar alolol ainol char oshd shar oshd
ochar ar yshd charoshd char dan chol oshdar ain shol shs
shal dain ain octhd char oshd daiin cholchol char oshd
shaldain yshal ochd chy ochar yshd dain cholshol chal an
shaldair yeear chol shol chardan qoshd yshdchar okain an
sholdain char am
From which we see 2 things - longer words in the TT generation tend to looser, and the algorithm used tends to pick up on words and run with them in a way the original scribe didn't (ie shaldain in the second paragraph)