5D Library › 玛雅卓尔金双模齿轮、神谕十字对合变换与六十四闸门网络渗流 · English

第三章 五大神谕:印记与音阶上的对合与群作用

四条规则

Dreamspell 为每个 kin 配上四个「神谕」位置:指引(guide)、支持(analog)、挑战(antipode)与隐藏推动(occult),与本身合称「五大神谕」 (Argüelles and Argüelles 1990)。把日符记为 s∈ℤ20s \in \mathbb{Z}_{20}(第 20 个日符记作 0),音阶记为 t∈{1,…,13}t \in \{1, \dots, 13\},通行的规则是:

  • 支持:A(s,t)=(19−s,t)A(s, t) = (19 – s,\ t),即两个日符之和模 20 为 19;
  • 挑战:P(s,t)=(s+10,t)P(s, t) = (s + 10,\ t),即日符相差 10;
  • 隐藏推动:O(s,t)=(21−s,14−t)O(s, t) = (21 – s,\ 14 – t),即日符之和为 21、音阶之和为 14;
  • 指引:日符保持同一颜色,按音阶偏移(第二节)。

颜色由 smod⁡4s \bmod 4 决定:1 红、2 白、3 蓝、0 黄。

推导

# Chapter 3. p04: The Dreamspell oracle as a group action on Z20 x Z13.
# Seals 1..20 are taken mod 20 (seal 20 = 0), tones 1..13.
#   analog   A(s, t) = (19 - s, t)       antipode P(s, t) = (s + 10, t)
#   occult   O(s, t) = (21 - s, 14 - t)
# Colour = s mod 4 (1 red, 2 white, 3 blue, 0 yellow). We check that A, P
# and O are involutions without fixed kins, find the group they generate
# on the seals, and how they act on the four colours.
from itertools import product

KINS = list(product(range(20), range(1, 14)))

def A(k):
    return ((19 - k[0]) % 20, k[1])

def P(k):
    return ((k[0] + 10) % 20, k[1])

def O(k):
    return ((21 - k[0]) % 20, 14 - k[1])

def closure(gens):
    ident = tuple(range(20))
    group, frontier = {ident}, [ident]
    while frontier:
        g = frontier.pop()
        for h in gens:
            gh = tuple(h[g[i]] for i in range(20))
            if gh not in group:
                group.add(gh)
                frontier.append(gh)
    return group

if __name__ == "__main__":
    for name, f in (("analog", A), ("antipode", P), ("occult", O)):
        inv = all(f(f(k)) == k for k in KINS)
        fixed = sum(f(k) == k for k in KINS)
        print("   {:8s} involution {}, fixed kins {}".format(
            name, inv, fixed))
    tone_fixed = [t for t in range(1, 14) if 14 - t == t]
    print("   occult fixes the tone only at t = {}".format(tone_fixed))
    gens = [tuple(A((s, 1))[0] for s in range(20)),
            tuple(P((s, 1))[0] for s in range(20)),
            tuple(O((s, 1))[0] for s in range(20))]
    G = closure(gens)
    orbit = {g[1] for g in G}
    print("   group generated on seals: order {}, orbit of seal 1: {}".format(
        len(G), len(orbit)))
    rot = tuple(A((O((s, 1))[0], 1))[0] for s in range(20))
    order, cur = 1, list(rot)
    while cur != list(range(20)):
        cur = [rot[x] for x in cur]
        order += 1
    print("   A after O is the shift s -> s + {} (order {}); P = shift 10"
          .format(rot[0], order))
    short = {1: "R", 2: "W", 3: "B", 0: "Y"}
    for name, f in (("analog", A), ("antipode", P), ("occult", O)):
        same = all(f((s, 1))[0] % 4 == f((s % 4, 1))[0] % 4
                   for s in range(20))
        m = " ".join("{}>{}".format(short[c], short[f((c, 1))[0] % 4])
                     for c in (1, 2, 3, 0))
        print("   {:8s} on colours: {}  (well defined: {})".format(
            name, m, same))
    ap = all(A(P((c, 1)))[0] % 4 == O((c, 1))[0] % 4 for c in range(4))
    print("   on colours analog o antipode = occult: {} (Klein four-group)"
          .format(ap))
#    analog   involution True, fixed kins 0
#    antipode involution True, fixed kins 0
#    occult   involution True, fixed kins 0
#    occult fixes the tone only at t = [7]
#    group generated on seals: order 20, orbit of seal 1: 20
#    A after O is the shift s -> s + 18 (order 10); P = shift 10
#    analog   on colours: R>W W>R B>Y Y>B  (well defined: True)
#    antipode on colours: R>B W>Y B>R Y>W  (well defined: True)
#    occult   on colours: R>Y W>B B>W Y>R  (well defined: True)
#    on colours analog o antipode = occult: True (Klein four-group)

三个对合与克莱因四元群

程序验证:AA、PP、OO 都是对合(作用两次回到原处),而且在全部 260 个 kin 上没有一个不动点。原因是代数的:19−s=s19 – s = s 与 21−s=s21 – s = s 要求 2s2s 为奇数,在 ℤ20\mathbb{Z}_{20} 中无解;s+10=ss + 10 = s 也无解。隐藏推动只在音阶 7 上保持音阶不变(14−7=714 – 7 = 7),但日符仍然改变。

在日符上,三个变换生成一个 20 阶的群,任一日符的轨道就是全部 20 个日符,所以这个作用是正则的:一个日符的全部「神谕关系」由群元唯一确定。AA 与 OO 是两个反射,复合 A∘OA \circ O 是平移 s↦s+18s \mapsto s + 18(阶为 10),而挑战 PP 恰是它的五次方,即平移 10;这正是 20 阶二面体群的结构。

在颜色上,三个变换各是一个没有不动点的对合:支持交换红白、蓝黄;挑战交换红蓝、白黄;隐藏推动交换红黄、白蓝;且「支持之后再挑战」等于「隐藏推动」。它们与恒等变换一起构成克莱因四元群 V4V_4,在四种颜色上正则作用。Dreamspell 所说的「颜色家族」的互补关系,就是这一个四元群。

指引是一个仿射映射

通行的指引表按音阶分为五类:音阶 1、6、11 的指引是自己的日符;2、7、12 为加 12;3、8、13 为加 4;4、9 为减 4;5、10 为减 12。

推导

# Chapter 3. p05: The guide as an affine map, and oracle ties between two
# people. Published guide table: tones 1, 6, 11 -> same seal; 2, 7, 12 ->
# +12; 3, 8, 13 -> +4; 4, 9 -> -4; 5, 10 -> -12 (seals mod 20). We check
# that this is the single rule G(s, t) = (s + 12 (t - 1) mod 20, t).
# Then, for two kins drawn independently and uniformly, how often is the
# second one of the first one's four oracle kins?
from itertools import product

TABLE = {1: 0, 6: 0, 11: 0, 2: 12, 7: 12, 12: 12, 3: 4, 8: 4, 13: 4,
         4: -4, 9: -4, 5: -12, 10: -12}
KINS = list(product(range(20), range(1, 14)))

def guide(k):
    return ((k[0] + 12 * (k[1] - 1)) % 20, k[1])

def oracle(k):
    s, t = k
    return {"guide": guide(k), "analog": ((19 - s) % 20, t),
            "antipode": ((s + 10) % 20, t), "occult": ((21 - s) % 20, 14 - t)}

if __name__ == "__main__":
    ok = all((s + TABLE[t]) % 20 == guide((s, t))[0] for s, t in KINS)
    print("   table = s + 12(t - 1) mod 20 for all 260 kins: {}".format(ok))
    selfg = sum(guide(k) == k for k in KINS)
    print("   kins that are their own guide: {} ({:.4f})".format(
        selfg, selfg / 260))
    same_colour = all(guide(k)[0] % 4 == k[0] % 4 for k in KINS)
    print("   guide keeps the colour: {}".format(same_colour))
    sizes = {}
    for k in KINS:
        n, cur = 1, guide(k)
        while cur != k:
            cur, n = guide(cur), n + 1
        sizes[n] = sizes.get(n, 0) + 1
    print("   iterating the guide, cycle length: kins {}".format(
        dict(sorted(sizes.items()))))
    hits = {"guide": 0, "analog": 0, "antipode": 0, "occult": 0}
    any_tie = 0
    for a, b in product(KINS, KINS):
        if a == b:
            continue
        o = oracle(a)
        tie = False
        for key, v in o.items():
            if v == b:
                hits[key] += 1
                tie = True
        any_tie += tie
    tot = 260 * 259
    for key, v in hits.items():
        print("   P(other is my {:8s}) = {:.5f}".format(key, v / tot))
    print("   P(other is in my oracle at all) = {:.5f}".format(any_tie / tot))
    print("   12 people: expected ordered oracle ties {:.2f}".format(
        12 * 11 * any_tie / tot))
#    table = s + 12(t - 1) mod 20 for all 260 kins: True
#    kins that are their own guide: 60 (0.2308)
#    guide keeps the colour: True
#    iterating the guide, cycle length: kins {1: 60, 5: 200}
#    P(other is my guide   ) = 0.00297
#    P(other is my analog  ) = 0.00386
#    P(other is my antipode) = 0.00386
#    P(other is my occult  ) = 0.00386
#    P(other is in my oracle at all) = 0.01455
#    12 people: expected ordered oracle ties 1.92

程序证明,这张表等价于一条公式:

G(s,t)=(s+12(t−1)mod⁡20,t).G(s, t) = \bigl(s + 12\,(t – 1) \bmod 20,\ t\bigr).

因为 12×5=60≡0(mod⁡20)12 \times 5 = 60 \equiv 0 \pmod{20},偏移只取决于 tmod⁡5t \bmod 5;又因为 12 是 4 的倍数,指引永远与本身同色。音阶为 1、6、11 的 60 个 kin(0.2308)以自己为指引;其余 200 个 kin 反复取指引时形成长度为 5 的循环。

两个人之间的「神谕关系」基率

若两个人的 kin 独立且均匀,一个人恰是另一个人的支持、挑战或隐藏推动的概率各为 0.00386,是其指引的概率为 0.00297,落入其四个神谕位置之一的概率合计 0.01455。在一个 12 人的团体中,按有序对计算,期望有 1.92 个这样的关系。也就是说,在任何稍大的团体里「发现」神谕关系几乎是必然的。

假说 · 尚未验证

作者的假说(尚未验证): 当事人报告的「与神谕伙伴特别投缘」的经验,其频率与团体人数成正比,正如基率所预言,而与两人是否真有神谕关系无关。什么会证明它错: 若在盲法条件下(当事人不知彼此的 kin),有神谕关系的配对报告投缘的比例显著高于无关系的配对,这一假说就不成立。第八章给出相应的样本量。

小结

支持、挑战、隐藏推动是 ℤ20×ℤ13\mathbb{Z}_{20} \times \mathbb{Z}_{13} 上没有不动点的对合,在日符上生成 20 阶的二面体群并正则作用,在颜色上构成克莱因四元群;指引是仿射映射 s+12(t−1)s + 12(t-1)。两个随机的人之间存在神谕关系的概率为 0.01455。

参考文献

Argüelles, José, and Lloydine Argüelles. 1990. Dreamspell: The Journey of Timeship Earth 2013. Boxed game kit.

Fan, Y. (2026). Mayan Tzolkin Modular Gears, Oracle-Cross Involutions and Sixty-Four-Gate Network Percolation: Coprime Cycle Synchronization, Antipode-Analog Permutations and Nine-Center Connected Components [玛雅卓尔金双模齿轮、神谕十字对合变换与六十四闸门网络渗流] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.24090193 · Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).