In conclusion, I would like to clarify once again that my model is based primarily (87.4%) on a table containing 485 entries (syllable combinations). The ledger words only fill in the gaps not covered by short_words and fixed words. They therefore account for only a small percentage, but they are necessary to ensure that all words are generated. Otherwise, the model as a whole would not function properly.
Proportions of word generation:
-----------------------------------------
Short words: 1634 (17.2%)
Common words: 4474 (47.1%)
Common combinations: 2195 (23.1%)
Ledger words: 1197 (12.6%)
Total: 9500
The high proportion of predefined words suggests that the author may indeed have worked with a tables book or combinatorics. While this would be labour-intensive, it would still be feasible for a 15th-century author.
Of course, that doesn't prove the manuscript is meaningless. A table book could also encode a real language or a cipher system.
generated.zip (Size: 20.11 KB / Downloads: 11)
Proportions of word generation:
-----------------------------------------
Short words: 1634 (17.2%)
Common words: 4474 (47.1%)
Common combinations: 2195 (23.1%)
Ledger words: 1197 (12.6%)
Total: 9500
The high proportion of predefined words suggests that the author may indeed have worked with a tables book or combinatorics. While this would be labour-intensive, it would still be feasible for a 15th-century author.
Of course, that doesn't prove the manuscript is meaningless. A table book could also encode a real language or a cipher system.
generated.zip (Size: 20.11 KB / Downloads: 11)