关于「东西方原典考辨:《士商类要》与开罗 Geniza 书信」,《商帮多边惩罚与买手信任网络》给出的定量结论是什么?

结论:《士商类要》代表的明代商书,与开罗 Geniza 中的马格里布商人书信,从两个方向回答了同一个问题。商书的规训告诉一位商人在消息不通的地方如何自保,p71 说明同乡网络、社会惩罚与担保层层叠加,可以把可托付的订单额放大二十多倍。Geniza 书信展示了一群商人如何通过通信交换声誉信息,p72 说明这种信息受季节航运的束缚,冬季的延迟足以让声誉的约束在几个月里松动。两部原典都没有告诉我们这些机制的边界在哪里,那是今天的数据可以回答的问题。最后一章要把全书的模型变成一份买手联盟可以照着执行的协议:名单怎么建、保证金怎么收、评分卡怎么算、什么样的结果会推翻本书?

数据口径:书中代理人、订单与网络参数均为示意;全部计算由书中程序生成;本书是非个人化的经营分析,不构成经营、法律或理财建议。 第四章求临界群规模:链状熟人网络在示意参数下只有几十到一百多人,均匀混合网络要到 33742 人。

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。

settingG_maxx strangerd* (G = 5)
stranger broker0.8181.00none
native-place news5.8977.210.8772
+ standing at home8.84510.810.8097
+ guarantor g=0.517.69021.620.6579
each layer multiplies the order a merchant can entrust
查看完整验算程序
# Chapter 7. p71: native-place networks, guarantors and the stranger.
# Model: G_max = a r / ((1 - a)(1 - g)), with a = d (1 - q),
# q = H (1 - s), r = w - w0 and g the share of a theft that a guarantor
# recovers from the agent or his kin. Four settings: a broker who is a
# stranger (no one else hears, s = 0); a native-place network in which
# news reaches s = 0.9 of members within the period; the same network
# where exclusion also costs the agent standing at home (w0 falls from
# 1 to 0.5); and the same with a guarantor recovering g = 0.5.
# d = 0.90, H = 0.5, w = 2 are illustrative, not drawn from sources.
# 模型:陌生牙人、同乡网络、同乡加社会惩罚、再加保人四种情形。

D, H, W = 0.90, 0.5, 2.0


def g_max(s, w0, g):
    a = D * (1.0 - H * (1.0 - s))
    return a * (W - w0) / ((1.0 - a) * (1.0 - g))


def d_star(gain, s, w0, g):
    # smallest d with gain (1 - g) <= d (1 - q)(gain (1 - g) + r)
    q = H * (1.0 - s)
    eff = gain * (1.0 - g)
    return eff / ((1.0 - q) * (eff + W - w0))


if __name__ == "__main__":
    cases = (("stranger broker", 0.0, 1.0, 0.0),
             ("native-place news", 0.9, 1.0, 0.0),
             ("+ standing at home", 0.9, 0.5, 0.0),
             ("+ guarantor g=0.5", 0.9, 0.5, 0.5))
    base = g_max(0.0, 1.0, 0.0)
    print("  setting               G_max    x stranger   d* (G = 5)")
    for name, s, w0, g in cases:
        gm = g_max(s, w0, g)
        ds = d_star(5.0, s, w0, g)
        dtxt = "{:.4f}".format(ds) if ds < 1.0 else "none"
        print("  {:<20s}  {:6.3f}   {:6.2f}       {}".format(
            name, gm, gm / base, dtxt))
    print("each layer multiplies the order a merchant can entrust")

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《商帮多边惩罚声誉机制、马格里布商人代理博弈与买手社群信任网络母典》第 7 章「东西方原典考辨:《士商类要》与开罗 Geniza 书信」(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-222 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026222,
  author    = {Fan, Yafu},
  title     = {Multilateral Punishment in Merchant Networks: Maghribi Agency Games and Trust Networks in Buyer Communities},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论