Sensitivity analysis with the second answers substituted (X5; exploratory, added after the results were seen, not part of the decision rule)
Method: in batches 3 and 6 of pass 1 the coder returned a second answer; under the rule fixed before the analysis (the first complete valid answer of each coder instance counts) it was not used.
Here the first answers for the new items of those two batches are replaced by the second answers; everything else stays as it was (the blind codes of the 90 first-sample items and pass 2 are untouched), and the main test and the related tests are recomputed.
The second answer in batch 3 was given by the same coder instance in the same context as the first, so it is not independent of the first; the second answer in batch 6 is a fresh run of the same prompt and is independent of the first (analysis plan, section 7).
The label X5 keeps this analysis apart from S6 of analysis plan section 4.3 (counting only Y).

1. How the two answers differ (new items, anchor items excluded)
  Batch 3 (97 new items): four categories (Y/E/N/X) identical in 73/97 (75.3%, kappa 0.551); three categories (related Y/E, unrelated N, proper name X) identical in 78/97 (80.4%, kappa 0.624); X versus non-X differs in 5 items
  Batch 6 (96 new items): four categories (Y/E/N/X) identical in 78/96 (81.2%, kappa 0.707); three categories (related Y/E, unrelated N, proper name X) identical in 81/96 (84.4%, kappa 0.740); X versus non-X differs in 3 items
  Of the new items in the two batches, 186 have Middle Chinese readings at both ends (labelled 45, ordinary 141); for 147 of them the first and second answers give the same four-category code.
  Changes from the first to the second answer (new items in the frame, by group; X = proper name):
    labelled (45 items): related -> unrelated 5, unrelated -> related 3
    ordinary (141 items): related -> unrelated 11, unrelated -> X 6, unrelated -> related 6, X -> unrelated 2

2. The table of the main test (in the frame, with Middle Chinese readings, not X; near among the pairs coded related)
  Reference (first answers, as analysed): 634 pairs in the frame, of which 220 coded related (labelled 141, ordinary 79); near: labelled 73/141, ordinary 23/79
  Substitution (second answers in batches 3 and 6): 630 pairs in the frame, of which 213 coded related (labelled 139, ordinary 74); near: labelled 71/139, ordinary 21/74
  Batch 3 only: 629 pairs in the frame, of which 220 coded related (labelled 142, ordinary 78); near: labelled 73/142, ordinary 22/78
  Batch 6 only: 635 pairs in the frame, of which 213 coded related (labelled 138, ordinary 75); near: labelled 71/138, ordinary 22/75

3. The tests (GEE clustered by phonetic, one-sided p; rule: p < .05 and upper confidence limit >= 2 supports C)
  P1 near ~ label | related
      Reference (first answers): 73/141 vs 23/79, OR 2.35 [1.43, 3.86], p1 0.00038; MH 2.01 [0.75, 5.39]; rule: supports C (same word / minimal derivation)
      Substitution (batches 3 and 6): 71/139 vs 21/74, OR 2.39 [1.38, 4.14], p1 0.00089; MH 2.06 [0.77, 5.51]; rule: supports C (same word / minimal derivation)
      Batch 3 only: 73/142 vs 22/78, OR 2.32 [1.42, 3.79], p1 0.00041; MH 1.76 [0.63, 4.90]; rule: supports C (same word / minimal derivation)
      Batch 6 only: 71/138 vs 22/75, OR 2.4 [1.39, 4.14], p1 0.00088; MH 2.31 [0.89, 5.98]; rule: supports C (same word / minimal derivation)
  S1 MC identical ~ label | related
      Reference (first answers): 54/141 vs 15/79, OR 2.57 [1.35, 4.87], p1 0.0019; MH 1.57 [0.57, 4.32]; rule: supports C (same word / minimal derivation)
      Substitution (batches 3 and 6): 52/139 vs 15/74, OR 2.28 [1.20, 4.32], p1 0.0058; MH 1.84 [0.64, 5.30]; rule: supports C (same word / minimal derivation)
      Batch 3 only: 54/142 vs 14/78, OR 2.68 [1.40, 5.12], p1 0.0014; MH 1.71 [0.58, 5.01]; rule: supports C (same word / minimal derivation)
      Batch 6 only: 52/138 vs 16/75, OR 2.18 [1.16, 4.11], p1 0.0078; MH 1.68 [0.62, 4.55]; rule: supports C (same word / minimal derivation)
  S2 near ~ label | related, paronomastic glosses removed
      Reference (first answers): 43/94 vs 20/73, OR 2.09 [1.18, 3.71], p1 0.0059; MH 1.36 [0.33, 5.66]; rule: supports C (same word / minimal derivation)
      Substitution (batches 3 and 6): 42/93 vs 18/68, OR 2.1 [1.13, 3.89], p1 0.0096; MH 1.26 [0.30, 5.33]; rule: supports C (same word / minimal derivation)
      Batch 3 only: 44/96 vs 19/72, OR 2.12 [1.19, 3.76], p1 0.0053; MH 1.16 [0.26, 5.07]; rule: supports C (same word / minimal derivation)
      Batch 6 only: 41/91 vs 19/69, OR 2.06 [1.11, 3.80], p1 0.011; MH 1.46 [0.36, 5.91]; rule: supports C (same word / minimal derivation)
  S3 near ~ label | related, labels shared with Xiao Xu only
      Reference (first answers): 41/93 vs 23/79, OR 1.81 [1.04, 3.15], p1 0.017; MH 1.24 [0.38, 4.07]; rule: supports C (same word / minimal derivation)
      Substitution (batches 3 and 6): 40/92 vs 21/74, OR 1.82 [1.01, 3.29], p1 0.023; MH 1.02 [0.29, 3.64]; rule: supports C (same word / minimal derivation)
      Batch 3 only: 42/95 vs 22/78, OR 1.84 [1.07, 3.17], p1 0.013; MH 0.95 [0.26, 3.46]; rule: supports C (same word / minimal derivation)
      Batch 6 only: 39/90 vs 22/75, OR 1.78 [0.98, 3.22], p1 0.029; MH 1.31 [0.41, 4.22]; rule: supports C (same word / minimal derivation)
  X3 near ~ label | related, shared and non-paronomastic (exploratory)
      Reference (first answers): 27/66 vs 20/73, OR 1.8 [0.95, 3.42], p1 0.036; MH 1.43 [0.29, 7.16]; rule: supports C (same word / minimal derivation)
      Substitution (batches 3 and 6): 26/65 vs 18/68, OR 1.79 [0.90, 3.60], p1 0.05 (exact one-sided p 0.0497, right at the .05 boundary); MH 1.3 [0.26, 6.63]; rule: supports C (same word / minimal derivation)
      Batch 3 only: 28/68 vs 19/72, OR 1.86 [0.98, 3.52], p1 0.029; MH 1.17 [0.22, 6.26]; rule: supports C (same word / minimal derivation)
      Batch 6 only: 25/63 vs 19/69, OR 1.72 [0.86, 3.44], p1 0.063; MH 1.57 [0.33, 7.50]; rule: indeterminate
  S6 near ~ label | coded Y only
      Reference (first answers): 24/38 vs 0/4, not estimable; rule: not estimable (GEE; e.g. complete separation)
      Substitution (batches 3 and 6): 20/33 vs 0/5, not estimable; MH inf [nan, nan]; rule: not estimable (GEE; e.g. complete separation)
      Batch 3 only: 23/36 vs 0/6, not estimable; MH inf [nan, nan]; rule: not estimable (GEE; e.g. complete separation)
      Batch 6 only: 21/35 vs 0/3, not estimable; rule: not estimable (GEE; e.g. complete separation)
  X1 near ~ label | coded E only (exploratory)
      Reference (first answers): 49/103 vs 23/75, OR 1.8 [1.00, 3.25], p1 0.025; MH 0.84 [0.27, 2.63]
      Substitution (batches 3 and 6): 51/106 vs 21/69, OR 1.89 [1.01, 3.55], p1 0.024; MH 1.28 [0.43, 3.81]
      Batch 3 only: 50/106 vs 22/72, OR 1.73 [0.96, 3.10], p1 0.034; MH 0.81 [0.23, 2.81]
      Batch 6 only: 50/103 vs 22/72, OR 1.99 [1.06, 3.71], p1 0.016; MH 1.27 [0.45, 3.61]
  S8 H2 related ~ label | all coded pairs, X excluded
      Reference (first answers): 141/172 vs 79/462, OR 21.57 [13.31, 34.98], p1 6.5e-36; MH 19.77 [10.10, 38.69]
      Substitution (batches 3 and 6): 139/172 vs 74/458, OR 21.96 [13.59, 35.48], p1 8.2e-37; MH 23.64 [11.83, 47.26]
      Batch 3 only: 142/172 vs 78/457, OR 22.75 [14.00, 36.98], p1 9.4e-37; MH 22.83 [11.56, 45.11]
      Batch 6 only: 138/172 vs 75/463, OR 20.81 [12.90, 33.56], p1 7.1e-36; MH 20.27 [10.26, 40.06]
  S9 near ~ related + label: coefficient of label
      Reference (first answers): 81/172 vs 83/462, OR 1.98 [1.30, 3.02], p1 0.00076
      Substitution (batches 3 and 6): 81/172 vs 83/458, OR 2.03 [1.29, 3.20], p1 0.0011
      Batch 3 only: 81/172 vs 82/457, OR 2.02 [1.30, 3.12], p1 0.0008
      Batch 6 only: 81/172 vs 84/463, OR 2.0 [1.29, 3.10], p1 0.001
  S9 near ~ related + label: coefficient of related
      Reference (first answers): 96/220 vs 68/414, OR 2.34 [1.51, 3.63], p1 7.8e-05
      Substitution (batches 3 and 6): 92/213 vs 72/417, OR 2.17 [1.38, 3.42], p1 0.0004
      Batch 3 only: 95/220 vs 68/409, OR 2.27 [1.45, 3.56], p1 0.00018
      Batch 6 only: 93/213 vs 72/422, OR 2.24 [1.44, 3.49], p1 0.00018
  X4 near ~ label | related, new items only (exploratory)
      Reference (first answers): 59/116 vs 21/73, OR 2.42 [1.37, 4.25], p1 0.0011; MH 1.26 [0.41, 3.90]
      Substitution (batches 3 and 6): 57/114 vs 19/68, OR 2.46 [1.33, 4.57], p1 0.0021; MH 1.38 [0.45, 4.26]
      Batch 3 only: 59/117 vs 20/72, OR 2.43 [1.37, 4.30], p1 0.0012; MH 1.12 [0.34, 3.69]
      Batch 6 only: 57/113 vs 20/69, OR 2.43 [1.33, 4.46], p1 0.0021; MH 1.51 [0.52, 4.43]

4. Comparison with an independent hand computation
  The main test with batches 3 and 6 replaced by the second answers, computed by hand by an independent reader: OR 2.39 [1.38, 4.14]. This script: 71/139 vs 21/74, OR 2.39 [1.38, 4.14], p1 0.00089; MH 2.06 [0.77, 5.51].
  The numbers agree.

Notes:
  1. This is an exploratory, post hoc analysis. The main test remains the first-answer result of analysis plan section 4.2 (OR 2.35); the rule fixed beforehand has not been changed.
  2. The substitution concerns only the new items of two batches (193 items, of which those in the frame are given in section 1); the blind codes of the 90 first-sample items, the second blind pass and the authors' check are untouched.
  3. The second answer in batch 3 comes from the same context as the first and is not an independent recoding; the second answer in batch 6 is an independent fresh run of the same prompt. The rows "Batch 3 only" and "Batch 6 only" show the contribution of each.
  4. The two answers were not merged into a "consensus" and no new coding was done.
