关于「原典考辨:韩愈论气与言、刘勰论四六、亚里士多德论节奏、朗吉努斯论和谐」,《机器文气与句长节律》给出的定量结论是什么?

结论:韩愈的「言之短长」、刘勰的「四字六字」、亚里士多德的「既非格律亦非无节奏」、朗吉努斯的「和谐与思想并重」,都在说同一件事:节律是内容的一部分。p71 显示,刘勰论句法的那一句整齐到变异系数只有 0.096,韩愈那一句则是两字与十二字的强烈对比;p72 显示,一篇名作的高变异系数可以只来自一个长句,抽样区间宽到不能当标准用。原典给了方向,没有给数值;数值要从台账里来。

数据口径:书中示例均为作者自写的合成文字或公有领域文本,参数为示意,全部计算由书中程序生成。 第一至六章依次讲句长指纹与对数正态、突发度与顺序量、检测器的贝叶斯账与偏差、套话密度的泊松检验与群体超额词汇、三元连发与压缩冗余、只动句读的动态规划与真实细节账本:合成草稿只改 5 处句读,变异系数就从 0.123 升到 0.546,而字词一个不动。

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: 0.096

Liu Xieclauses [6, 5, 6, 6, 6, 7]mean 6.00CV 0.096contrast 1.4
Han Yuclauses [2, 12]mean 7.00CV 0.714contrast 6.0
Liu Xie clauses within 3 to 7 hanzi: 6 of 6
查看完整验算程序
# Chapter 7. p71: clause lengths in two classical Chinese sentences.
# Texts as transcribed on zh.wikisource (simplified here): Liu Xie,
# Wenxin Diaolong, "Zhangju"; Han Yu, "Da Li Yi shu" (the Quan Tang wen
# reading, without the particle zhe). One sentence from each, public
# domain. Clauses are split at the full-width comma and semicolon;
# lengths are hanzi counts. The contrast ratio is longest / shortest.
import math
import re

LIU = "若夫笔句无常,而字有条数,四字密而不促,六字格而非缓,或变之以三五,盖应机之权节也。"
HAN = "气盛,则言之短长与声之高下皆宜。"


def clauses(s):
    parts = re.split(r"[,;。]", s)
    return [len(re.findall(r"[一-鿿]", p)) for p in parts if p]


for name, s in (("Liu Xie", LIU), ("Han Yu", HAN)):
    xs = clauses(s)
    m = math.fsum(xs) / len(xs)
    sd = math.sqrt(math.fsum((x - m) ** 2 for x in xs) / len(xs))
    print("{:<8} clauses {}  mean {:.2f}  CV {:.3f}  contrast {:.1f}".format(
        name, xs, m, sd / m, max(xs) / min(xs)))
# Liu Xie's claim read as a band: clause lengths 4 to 6, varied by 3 or
# 5. Share of his own clauses that fall inside 3 to 7 hanzi:
xs = clauses(LIU)
inside = len([x for x in xs if 3 <= x <= 7])
print("Liu Xie clauses within 3 to 7 hanzi: {} of {}".format(inside, len(xs)))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《机器生成文本统计指纹剥离、句长突发律动重构与宗师级文气呼吸感母典》第 7 章「原典考辨:韩愈论气与言、刘勰论四六、亚里士多德论节奏、朗吉努斯论和谐」(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-394 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026394,
  author    = {Fan, Yafu},
  title     = {Stripping Statistical Fingerprints of Machine Prose, Sentence-Length Burstiness Reconstruction and Master-Level Stylistic Cadence},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论