关于「东西方原典考辨:《尔雅》的类聚与夸美纽斯的《世界图解》」,《词汇覆盖率与速成边界》给出的定量结论是什么?

结论:《尔雅》以常用词训释罕见同义词,是按意义归并的古代词族;《世界图解》以图示物、按主题组织课程,降低了获得意义的成本。p71 表明,在一般文字上按主题整块收词的词表远不如频率词表,只有当文本高度集中于某个主题(主题比例达到 0.5 以上)时,混合编排才胜出;p72 表明图画与词典改变的是每个生词的价格,而不是生词的数量。两部原典都在提醒我们,覆盖率之外还有意义与情境;而覆盖率提醒它们,先学什么由频率决定。最后一章把全书的计算收拢成一套可以执行的流程:怎样为马来西亚与台湾的客服团队建词表、测覆盖率,又用什么样的预注册检验去推翻本书自己?

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。

English version

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: 0.5

budgetlistgeneralown-field textother-field text
981thematic0.35910.54830.2514
981frequency0.72210.70590.6891
981hybrid0.55670.68660.5110
1739thematic0.40150.57800.2811
1739frequency0.83930.83440.8234
1739hybrid0.71490.79730.6814
4457thematic0.55830.68770.3908
4457frequency0.94930.94620.9431
4457hybrid0.90020.92700.8732
own-field share 0.3: hybrid 0.7973, frequency 0.8344
own-field share 0.5: hybrid 0.8523, frequency 0.8311
own-field share 0.7: hybrid 0.9072, frequency 0.8279
own-field share 0.9: hybrid 0.9622, frequency 0.8246
查看完整验算程序
# Chapter 7. p71: a picture-book list by themes against a frequency list
# Orbis Pictus and the Erya arrange words by subject (heaven, earth,
# plants, tools, kin), not by frequency. Model: the family curve of p23
# (V = 20000, s = 2.389, q = 664.5); the top CORE = 200 families are
# grammar words that both lists contain; every other family is put in
# one of K = 30 subject fields by a seeded draw. A thematic list takes
# whole fields (all members up to general rank 12000) in a fixed order
# until its budget is spent; a frequency list takes the top families.
# A hybrid list takes the top half by frequency, then whole fields.
# A text "about" a field draws PHI = 0.3 of its tokens from that
# field's words. K, PHI and the rank cap are illustrative.
import math
import random

V, S, Q, CORE, K, CAP, PHI = 20000, 2.389, 664.5, 200, 30, 12000, 0.3


def setup():
    rng = random.Random(20261113)
    raw = [(j + Q) ** (-S) for j in range(1, V + 1)]
    tot = math.fsum(raw)
    freq = [x / tot for x in raw]
    field = [-1] * CORE + [rng.randrange(K) for _ in range(V - CORE)]
    return freq, field


def thematic(field, budget, base=CORE):
    chosen, fields = set(range(base)), []
    for f in range(K):
        members = [i for i in range(base, CAP) if field[i] == f]
        if len(chosen) + len(members) > budget:
            break
        chosen.update(members)
        fields.append(f)
    return chosen, fields


def in_field(freq, field, known, f, phi=PHI):
    mem = [i for i in range(V) if field[i] == f]
    fsum_f = math.fsum([freq[i] for i in mem])
    hit_f = math.fsum([freq[i] for i in mem if i in known]) / fsum_f
    hit_g = math.fsum([freq[i] for i in known])
    return math.fsum([phi * hit_f, (1 - phi) * hit_g])


if __name__ == "__main__":
    freq, field = setup()
    print("budget  list       general  own-field text  other-field text")
    for budget in (1000, 2000, 4500):
        th, fields = thematic(field, budget)
        fr = set(range(len(th)))
        other = [f for f in range(K) if f not in fields][0]
        own = fields[0]
        hy, hf = thematic(field, len(th), base=len(th) // 2)
        assert own in hf and len(hy) <= len(th)
        for name, known in (("thematic", th), ("frequency", fr),
                            ("hybrid", hy)):
            print("  {:4d}  {:9s}  {:.4f}   {:.4f}          {:.4f}".format(
                len(th), name, math.fsum([freq[i] for i in known]),
                in_field(freq, field, known, own),
                in_field(freq, field, known, other)))
    th, fields = thematic(field, 2000)
    hy, hf = thematic(field, len(th), base=len(th) // 2)
    fr = set(range(len(th)))
    for phi in (0.3, 0.5, 0.7, 0.9):
        a = in_field(freq, field, hy, fields[0], phi)
        b = in_field(freq, field, fr, fields[0], phi)
        print("own-field share {:.1f}: hybrid {:.4f}, frequency {:.4f}"
              .format(phi, a, b))

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 7 章「东西方原典考辨:《尔雅》的类聚与夸美纽斯的《世界图解》」(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}
}

本书其他章节结论