关于「东西方原典考辨:刘向校雠与巴特利特的串行复现」,《衍生品的出处与保真》给出的定量结论是什么?
结论:第七章考辨刘向《别录》佚文「一人持本,一人读书,若怨家相对」与巴特利特的串行复现:两人独立对校剩下 0.12 的错误。
数据口径:书中参数为示意,代价用记账点数而非价格;全部计算由书中程序生成。 书中以标注框区分已发表的证据、可运行的 Python 计算(19 段程序)、原典与作者的假说。
对照数据表
下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.12
| solo reader A leaves 0.30; pair leaves 0.12 of the errors | ||||||
| case | n1 | n2 | both | found | LP-est | Chapman |
| homogeneous | 42.03 | 36.07 | 25.30 | 52.80 | 60.22 | 59.92 |
| half hard | 29.91 | 27.06 | 15.30 | 41.67 | 53.70 | 52.86 |
| true total in every document: 60 |
查看完整验算程序
# Chapter 7. p71: one reader (jiao) against two readers facing off (chou).
# A derivative holds T = 60 discrepancies from the master (illustrative).
# Reader A catches each with probability 0.7, reader B with 0.6. One
# reader alone leaves 1 - 0.7 of them; two independent readers leave
# (1 - 0.7)(1 - 0.6). The overlap m of their finds also estimates the
# total: Lincoln-Petersen n1 n2 / m and Chapman's correction. If half
# the discrepancies are hard (each reader catches only 0.3 of those),
# misses are correlated and the estimate falls short.
# Fixed-seed Monte Carlo, 5000 documents per case.
import math
import random
T, RUNS = 60, 5000
def trial(rng, pa, pb, hard_share, hard_p):
n1 = n2 = m = 0
for i in range(T):
hard = i < int(T * hard_share)
a = rng.random() < (hard_p if hard else pa)
b = rng.random() < (hard_p if hard else pb)
n1 += a
n2 += b
m += a and b
return n1, n2, m
rng = random.Random(20261113)
print("solo reader A leaves %.2f; pair leaves %.2f of the errors"
% (1 - 0.7, (1 - 0.7) * (1 - 0.6)))
print("%-14s %6s %6s %6s %8s %8s %8s" % ("case", "n1", "n2", "both",
"found", "LP-est", "Chapman"))
for name, hs in (("homogeneous", 0.0), ("half hard", 0.5)):
rows = [trial(rng, 0.7, 0.6, hs, 0.3) for _ in range(RUNS)]
n1 = math.fsum(r[0] for r in rows) / RUNS
n2 = math.fsum(r[1] for r in rows) / RUNS
m = math.fsum(r[2] for r in rows) / RUNS
found = n1 + n2 - m
lp = math.fsum(r[0] * r[1] / r[2] for r in rows if r[2]) / RUNS
ch = math.fsum((r[0] + 1) * (r[1] + 1) / (r[2] + 1) - 1
for r in rows) / RUNS
print("%-14s %6.2f %6.2f %6.2f %8.2f %8.2f %8.2f" % (
name, n1, n2, m, found, lp, ch))
print("true total in every document: %d" % T)唯一数据源与引用
范雅福(Fan, Y.). (2026). 《母书多模态衍生编译与版权链、单部专著到播客投影片题库的结构同态流水线母典》第 7 章「东西方原典考辨:刘向校雠与巴特利特的串行复现」(版本 1.0.0). 5D Library. DOI 待发布
DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-268 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026268,
author = {Fan, Yafu},
title = {Multimodal Derivative Compilation and Provenance Chains: A Structure-Preserving Pipeline from One Monograph to Podcasts, Slides and Quiz Banks},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}