关于「东西方原典考辨:「性相近,习相远」与亚里士多德的习性」,《习惯渐近曲线与二十一天神话》给出的定量结论是什么?
结论:《论语》说「性相近也,习相远也」,模型显示:在同一件事上,速率差别让人在第 21 天前后拉开最大距离,标准差 5.68 分,一年后收窄到 0.16 分;做与不做的差别则持续扩大到 14.99 分。亚里士多德说习惯完成自然所赋予的能力,这与平台和速率的区分一致;他的石头即使抛一万次也不会向上,正如平台为零的曲线永远是零。「从小养成关系到一切」在模型中有上限:旧习惯保持半年以上,新习惯都在第 40 天前后超过它。最后一章把全书的结论变成一个可以执行、也可以被推翻的实证闭环。
数据口径:书中的人群参数、遗忘速率与线索指数均为示意;全部计算由书中程序生成;本书不是医疗或心理建议,也不是个人化的经营建议。 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: 0.16, 14.99, 21, 5.68
| day | sd world A | sd world B | sd world C |
|---|---|---|---|
| 0 | 1.00 | 1.00 | 1.00 |
| 10 | 4.92 | 6.62 | 1.30 |
| 21 | 5.68 | 9.82 | 1.68 |
| 42 | 4.74 | 12.49 | 1.71 |
| 66 | 3.42 | 13.71 | 1.37 |
| 120 | 1.66 | 14.61 | 0.78 |
| 365 | 0.16 | 14.99 | 0.54 |
| world A: widest spread 5.68 points on day 21 of those shown |
查看完整验算程序
# Chapter 7. p71: "By nature near, by practice far" as a variance curve.
# 2000 people start alike (start 5, sd 1 on the 0-42 scale). Three worlds:
# (A) all practise the same act daily, rates differ (T95 lognormal, median
# 66, sigma 0.593 from p11): spread first opens, then closes;
# (B) half practise the act, half never do: spread opens and stays;
# (C) all practise, compliance p spread evenly from 0.6 to 1.0, with the
# expected-value form of the p41 model (step k, slide k / 10).
# Starts are shuffled (seeded) so they are independent of rate and
# compliance. We print the between-person sd by day. Illustrative.
import math
import random
import statistics
from statistics import NormalDist
Z = NormalDist()
N, A = 2000, 35.0
L20 = math.log(20 * 30 / 35)
zs = [Z.inv_cdf((i + 0.5) / N) for i in range(N)]
starts = [5.0 + z for z in zs]
random.Random(20261112).shuffle(starts) # start independent of rate
rates = [L20 / (66.0 * math.exp(0.593 * z)) for z in zs]
K = 1.0 - math.exp(-L20 / 66.0)
D = K / 10.0
ps = [0.6 + 0.4 * (i + 0.5) / N for i in range(N)]
def world_a(t):
return [A - (A - s) * math.exp(-c * t) for s, c in zip(starts, rates)]
def world_b(t):
return [A - (A - s) * math.exp(-c * t) if i % 2 == 0 else s
for i, (s, c) in enumerate(zip(starts, rates))]
def world_c(t):
out = []
for s, p in zip(starts, ps):
x = s
for _ in range(t):
x += p * K * (A - x) - (1 - p) * D * (x - s)
out.append(x)
return out
print(" day sd world A sd world B sd world C")
peak = (0.0, 0)
for t in (0, 10, 21, 42, 66, 120, 365):
sa = statistics.pstdev(world_a(t))
sb = statistics.pstdev(world_b(t))
sc = statistics.pstdev(world_c(t))
peak = max(peak, (sa, t))
print(" %3d %10.2f %10.2f %10.2f" % (t, sa, sb, sc))
print("")
print("world A: widest spread %.2f points on day %d of those shown"
% peak)唯一数据源与引用
范雅福(Fan, Y.). (2026). 《习惯形成渐近曲线、二十一天神话证伪与情境线索自动化指数母典》第 7 章「东西方原典考辨:「性相近,习相远」与亚里士多德的习性」(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-197 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026197,
author = {Fan, Yafu},
title = {Asymptotic Habit-Formation Curves, Debunking the 21-Day Myth and a Cue-Based Automaticity Index},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}