关于「年龄峰:气味把人带回哪一段人生」,《声香签名与气味记忆》给出的定量结论是什么?
结论:气味线索把自传记忆的分布重心向前推了约十年,峰值落在十岁以前,词语线索的峰值则在青少年到成年早期。这是分布的平移,不是开关:近期记忆在两种线索下同样多。程序 p22 显示,双峰分布下中位数毫无用处,占比才是合适的指标;程序 p23 显示,组间差距越小,所需样本按平方增长。品牌层面,香氛帮助记忆的路径更可能是注意,而不是唤起往事。
数据口径:书中参数为示意,全部计算由书中程序生成。 本书不是医疗、心理咨询或法规意见,不宣称任何身心功效。
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
| age bin | odour | verbal | odour bar |
|---|---|---|---|
| 0-4 | 0.114 | 0.008 | ########### |
| 5-9 | 0.259 | 0.042 | ########################## |
| 10-14 | 0.150 | 0.120 | ############### |
| 15-19 | 0.022 | 0.178 | ## |
| 20-24 | 0.001 | 0.137 | |
| 25-29 | 0.002 | 0.056 | |
| 30-34 | 0.004 | 0.015 | |
| 35-39 | 0.009 | 0.010 | # |
| 40-44 | 0.021 | 0.021 | ## |
| 45-49 | 0.049 | 0.048 | ##### |
| 50-54 | 0.112 | 0.111 | ########### |
| 55-59 | 0.258 | 0.255 | ########################## |
| odour | bump bin 5-9, before 10: 0.373, last 10 y: 0.370 | ||
| verbal bump bin 15-19, before 10: 0.050, last 10 y: 0.365 |
查看完整验算程序
# Chapter 2. p21: where cued autobiographical memories come from in life.
# Illustrative model for a 60-year-old. The age at encoding of a
# retrieved memory has density proportional to a bump term plus a
# recency term: w(a) = b N(a; mu, s) + (1 - b) exp(-(60 - a) / 6) / 6,
# with bump weight b = 0.55 for both cues.
# Odour cue: bump at mu = 8, s = 4 (direction from Chu and Downes 2000,
# peak 6 to 10, and Willander and Larsson 2006, first decade).
# Verbal cue: bump at mu = 18, s = 6 (peak 11 to 25 in the same work).
# Bin widths are 5 years; numbers are a sketch, not fitted data.
# The bump bin is the peak among bins before age 40 (the recency
# tail of the last decade is reported separately).
import math
from statistics import NormalDist
AGE = 60
BINS = list(range(0, AGE, 5))
CUES = {"odour": (8.0, 4.0, 0.55), "verbal": (18.0, 6.0, 0.55)}
def density(a, mu, s, b):
bump = NormalDist(mu, s).pdf(a)
rec = math.exp(-(AGE - a) / 6.0) / 6.0
return b * bump + (1.0 - b) * rec
def shares(cue):
mu, s, b = CUES[cue]
raw = []
for lo in BINS:
pts = [density(lo + 0.25 + 0.5 * k, mu, s, b) for k in range(10)]
raw.append(math.fsum(pts) * 0.5)
tot = math.fsum(raw)
return [x / tot for x in raw]
if __name__ == "__main__":
so, sv = shares("odour"), shares("verbal")
print("age bin odour verbal odour bar")
for lo, a, b in zip(BINS, so, sv):
print(" {:>2}-{:<2} {:.3f} {:.3f} {}".format(
lo, lo + 4, a, b, "#" * int(round(a * 100))))
for name, s in (("odour", so), ("verbal", sv)):
head = s[:8]
peak = BINS[head.index(max(head))]
print("{:<6} bump bin {}-{}, before 10: {:.3f}, last 10 y: {:.3f}"
.format(name, peak, peak + 4, math.fsum(s[:2]),
math.fsum(s[-2:])))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《声学节律与嗅觉分子跨模态耦合、普鲁斯特记忆唤醒与瞬时状态锚定母典》第 2 章「年龄峰:气味把人带回哪一段人生」(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-392 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026392,
author = {Fan, Yafu},
title = {Cross-Modal Coupling of Acoustic Rhythms and Olfactory Cues, Proustian Memory Retrieval and Rapid State-Conditioning Anchors},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}