coded pairs (labelled, ordinary, all): (212, 555, 767)
new items, first-sample items in the frame: (677, 90)
labelled without an MC reading: 35
labelled proper names with an MC reading: 5
ordinary proper names: 93
pairs in the tests (labelled, ordinary): (172, 462)
phonetics in the tests: 150
disagreements (all, labelled, ordinary): (76, 33, 43)
share of the 677: 11.2
labelled Y/E, E/N: (16, 15)
ordinary E/N, N/X: (30, 10)
related on one pass, N on the other (labelled, ordinary): (17, 31)
new items per group (labelled, ordinary): (177, 500)
author-check figures found in ext_author_check_output.txt
Table 3 examples: both passes agree with confidence 3
Table 5 rows: 16
Table 5: with a paronomastic gloss: 8
Table 5: shared by both recensions: 11
Table 5: layer undecidable: ['緉']
Table 5: blind codes (Y, E, X): (8, 7, 1)
homophonous labelled pairs by type: {'F': 42, 'L': 8, 'U': 8, 'C': 5}
homophonous ordinary pairs by type (172-phonetic frame): {'U': 73, 'X': 42, 'F': 15, 'L': 3}
homophonous ordinary pairs (frame, all rows): (133, 143)
141 x 23/79: 41
Section 4.4 illustrative pairs: reading categories and codes as in the text
Wang Yun's nine labels by reading category: {'alt_departing': 3, 'alt_other': 1, 'identical': 4, 'other': 1}
212 entries by recension layer (both, Da Xu only, undecidable): {'both': 140, 'daxu_only': 62, 'unknown': 10}
three of Wang Yun's nine labels present in both recensions: ['both', 'both', 'both']
labels filed under their own phonetic (all, in both recensions): (40, 37)
the 62 entries only in Da Xu include 喪, which Xiao Xu reads as a plain phonetic compound: [['daxu_only', '聲']]
all counts agree with the files
