关于「百分之九十八:覆盖率阈值的原始证据与它的边界」,《词汇覆盖率与速成边界》给出的定量结论是什么?
结论:第一章把覆盖率译成纸面上的生词密度,并计算断崖与斜坡两种剂量—反应曲线的效应量;从九成五到九成八,每个覆盖点约需 1166.7 个词族,是前一段价格的 7.0 倍,从两千词族到九成八覆盖,词汇量要乘以 4.0 到 4.5。
数据口径:书中参数为示意,全部计算由书中程序生成;本书是非个人化的经营与教学参考,不是针对任何个人或企业的经营建议,也不承诺任何课程能在某个期限内带来流利。 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: 1166.7, 4.0, 4.5, 7.0, 95, 98
| step families points per point | |
| 2000 (80 %) | 4000 (95 %) 2000 15 133.3 |
| 2000 (80 %) | 4500 (95 %) 2500 15 166.7 |
| 2000 (80 %) | 5000 (95 %) 3000 15 200.0 |
| last three points, 95 % | 98 %: |
| 4000 | 8000 (L&RK ) 1333.3 per point, x 8.0 the first step |
| 4000 | 8500 (Nation mid ) 1500.0 per point, x 9.0 the first step |
| 4000 | 9000 (Nation high) 1666.7 per point, x 10.0 the first step |
| 4500 | 8000 (L&RK ) 1166.7 per point, x 7.0 the first step |
| 4500 | 8500 (Nation mid ) 1333.3 per point, x 8.0 the first step |
| 4500 | 9000 (Nation high) 1500.0 per point, x 9.0 the first step |
| 5000 | 8000 (L&RK ) 1000.0 per point, x 6.0 the first step |
| 5000 | 8500 (Nation mid ) 1166.7 per point, x 7.0 the first step |
| 5000 | 9000 (Nation high) 1333.3 per point, x 8.0 the first step |
| vocabulary multiple from 2000 families to 98 %: x 4.0 to x 4.5 |
查看完整验算程序
# Chapter 1. p13: families per coverage point along the published ladder
# Anchors (word families, coverage incl. proper nouns):
# 2000 -> 80 % the round figure quoted in the project blueprint
# 4000 to 5000 -> 95 % minimal threshold, Laufer and
# Ravenhorst-Kalovski (2010)
# 8000 -> 98 % optimal threshold, same paper
# 8000 to 9000 -> 98 % for written text, Nation (2006)
# For each step we print how many families buy one coverage point and
# how that price compares with the first step. Midpoints are marked.
from fractions import Fraction
LADDER = [("blueprint", 2000, 80), ("L&RK low", 4000, 95),
("L&RK mid", 4500, 95), ("L&RK high", 5000, 95)]
TOP = [("L&RK", 8000, 98), ("Nation mid", 8500, 98),
("Nation high", 9000, 98)]
def price(a, b):
return Fraction(b[1] - a[1], b[2] - a[2])
if __name__ == "__main__":
base = LADDER[0]
print("step families points per point")
for rung in LADDER[1:]:
p = price(base, rung)
print(" 2000 (80 %) -> {:5d} (95 %) {:8d} {:6d} {:9.1f}".format(
rung[1], rung[1] - 2000, 15, float(p)))
print("last three points, 95 % -> 98 %:")
first = price(base, LADDER[2])
for low in LADDER[1:]:
for top in TOP:
p = price(low, top)
print(" {:5d} -> {:5d} ({:11s}) {:6.1f} per point,"
" x {:4.1f} the first step".format(
low[1], top[1], top[0], float(p),
float(p / first)))
print("vocabulary multiple from 2000 families to 98 %:"
" x {:.1f} to x {:.1f}".format(8000 / 2000, 9000 / 2000))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《二语词汇覆盖率九十八阈值、齐普夫频率曲线与马来语闽南语速成边界母典》第 1 章「百分之九十八:覆盖率阈值的原始证据与它的边界」(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-228 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026228,
author = {Fan, Yafu},
title = {The 98% Lexical-Coverage Threshold, Zipfian Frequency Curves and Fast-Track Limits for Malay and Hokkien Learners},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}