关于「《氓》《谷风》与《白头吟》」,《分手之后的记忆、信号与复合》给出的定量结论是什么?
结论:河有岸,湿地有边:凡事有界限。p13 的 (c) 部分把再试一次的决定写成期望值。如果对方改变了,复合值 10;没有改变,值 -6;那么只有当「他已经改变」的信念超过 0.375 时,再试一次才值得。信念为 0.200 时,再试一次的价值是 -2.80;为 0.600 时是 +3.60。《氓》的叙述者已经看够了,把信念放得很低,她的「亦已焉哉」正是这条决策规则在起作用。
数据口径:书中的参数均为示意或合成;全部计算由书中程序生成;本书不是心理咨询或治疗,也不是挽回任何人的操作手册,对方的决定属于对方。 第一章把复合当作带竞争事件的生存分析:在合成队列中,Kaplan-Meier 估计年底复合率为 0.456,真正复合的比例只有 0.311;并说明即使每对伴侣自身的机会不变,群体风险率也会因筛选降到最初的 0.355。
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: -2.80, 0.200, 0.375, 0.600, 10, 3.60
| (a) months memory of past care left (tp | 75 days) |
| 1 | 0.6668 |
| 3 | 0.2964 |
| 12 | 0.0077 |
| 36 | 0.0000 |
| (b) loss L | 10 p = 0.1 p = 0.5 p = 0.9 |
| gain g | 1 not deterred deterred |
| gain g | 4 not deterred deterred |
| gain g | 8 not not deterred |
| (c) break-even belief that he has changed | 0.375 |
| belief 0.200: value of trying again -2.80 | |
| belief 0.375: value of trying again +0.00 | |
| belief 0.600: value of trying again +3.60 |
查看完整验算程序
# Chapter 7. p13: Three old poems as arithmetic.
# (a) "Gu feng": "you do not think of former days". If the memory of a
# partner's past care fades with tp = 75 days unless it is renewed,
# how much is left after 1, 3 and 12 months, and after three years?
# (b) "Bai tou yin": the wife's threat to leave stops the husband from
# taking a concubine if his gain g from the concubine is smaller than
# p L, where L is his loss if she leaves and p the chance she really
# leaves. A threat she would not carry out has p near 0; a credible
# one has p near 1. Which g are deterred?
# (c) "Mang": "had I not thought it would turn, then let it end". A
# reunion is worth 10 if the partner has changed and -6 if not; with
# belief q that he has changed, trying again is worth 10 q - 6 (1 - q).
# The break-even belief is 6 / 16.
from math import exp
if __name__ == "__main__":
print(" (a) months memory of past care left (tp = 75 days)")
for m in (1, 3, 12, 36):
print(" {:10d} {:7.4f}".format(m, exp(-m * 30.4 / 75)))
print(" (b) loss L = 10 p = 0.1 p = 0.5 p = 0.9")
for g in (1, 4, 8):
row = ["deterred" if g < p * 10 else " not" for p in (0.1, 0.5,
0.9)]
print(" gain g = {:<2d} {:>8} {:>8} {:>8}".format(
g, *row))
print(" (c) break-even belief that he has changed: {:.3f}".format(
6 / 16))
for q in (0.2, 0.375, 0.6):
print(" belief {:.3f}: value of trying again {:+.2f}".format(
q, 10 * q - 6 * (1 - q)))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《关系断裂生存分析、断联期记忆衰减与复联博弈窗口》第 7 章「《氓》《谷风》与《白头吟》」(版本 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.24090241
DOI: 10.5281/zenodo.24090241 · 阅读本章所在专著 · 5DL-B-092 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026092,
author = {Fan, Yafu},
title = {Relationship Rupture Survival Analysis, No-Contact Memory Decay and Reconciliation Game Windows},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
doi = {10.5281/zenodo.24090241},
url = {https://doi.org/10.5281/zenodo.24090241}
}