关于「声音、笃定与校准:笃定何时是资产,何时是负债」,《施为言语与聚点定框》给出的定量结论是什么?

结论:第六章说明笃定只在校准时是资产:说「肯定」而出错,听者认为你是校准型的概率跌到 0.085,说「大概」而出错仍为 0.833;在完整报价里,言语承载 0.914 的决策信息,「7-38-55」不是沟通法则。

数据口径:书中对话全部合成,参数为示意,全部计算由书中程序生成;本书是非个人化经营建议,不是法律意见;分润只写单层、只给直接推荐人。 第一章用奥斯汀与塞尔的施为句理论,把报价拆成六个适切条件:示意模型里结构化报价把适切概率从 0.305 提到 0.734,一份完全可信的报价让对方期望熵下降 0.918 比特。

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: 0.085, 0.833

single checked claim P(calibrated) trust next 'certain'
certain right0.430 0.807
certain wrong0.085 0.721
probably right0.833 0.908
probably wrong0.833 0.908
4 right then 1 wrong, all 'certain': P(cal) 0.029
hedged where unsure, same miss rate: P(cal) 0.986
查看完整验算程序
# Chapter 6. p61: the calibration penalty as a Bayesian update.
# Two speaker types. Calibrated: says "certain" half the time and is
# then right 0.95 of the time; says "probably" otherwise and is right
# 0.70. Overconfident: says "certain" 0.9 of the time, "probably" 0.1,
# and is right 0.70 either way. A listener starts at P(calibrated) =
# 0.5 and, after each claim checked against the facts, updates. Her
# trust in the next "certain" claim is P(cal) * 0.95 + P(over) * 0.70.
# The sizes are illustrative; the direction follows Tenney et al. 2007.
from fractions import Fraction as F

CAL = {"certain": (F(1, 2), F(95, 100)), "probably": (F(1, 2), F(7, 10))}
OVER = {"certain": (F(9, 10), F(7, 10)), "probably": (F(1, 10), F(7, 10))}


def like(model, say, right):
    p_say, acc = model[say]
    return p_say * (acc if right else 1 - acc)


def update(prior, say, right):
    a = prior * like(CAL, say, right)
    b = (1 - prior) * like(OVER, say, right)
    return a / (a + b)


def trust(pc):
    return pc * F(95, 100) + (1 - pc) * F(7, 10)


print("single checked claim      P(calibrated)  trust next 'certain'")
for say in ("certain", "probably"):
    for right in (True, False):
        pc = update(F(1, 2), say, right)
        print("  %-8s %-5s           %.3f          %.3f" % (
            say, "right" if right else "wrong", pc, trust(pc)))
seq = [("certain", True)] * 4 + [("certain", False)]
pc = F(1, 2)
for s, r in seq:
    pc = update(pc, s, r)
print("4 right then 1 wrong, all 'certain': P(cal) %.3f" % pc)
seq2 = [("probably", True)] * 2 + [("certain", True)] * 2 + \
    [("probably", False)]
pc2 = F(1, 2)
for s, r in seq2:
    pc2 = update(pc2, s, r)
print("hedged where unsure, same miss rate:  P(cal) %.3f" % pc2)

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《施为性言语行为、谢林聚点定框与高维语言坍缩商业现实母典》第 6 章「声音、笃定与校准:笃定何时是资产,何时是负债」(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-428 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026428,
  author    = {Fan, Yafu},
  title     = {Performative Speech Acts, Schelling Focal-Point Framing and Verbal Reality-Collapse in High-Stakes Negotiation},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论