What does The Limits of Guardrail Compression conclude, in numbers, about “Classical Sources East and West: "Writing Does Not Exhaust Speech, Speech Does Not Exhaust Intent" and Shannon's Suspension of Meaning”?
Answer: Chapter 7 examines "writing does not exhaust speech, speech does not exhaust intent" from the Xici commentary, Wang Bi's "forget the image once the intent is grasped", and Shannon's decision to set meaning outside the engineering problem: with the same three rules, compressing to reproduce the draft keeps 0.6108 bits and compressing to keep the verdict 0.7234 bits.
Scope of the numbers: Chapter 1 builds an illustrative source of eight kinds of draft and three verdicts and the counting bound: three rules keep 0.960 of the verdict information, two only 0.629. Chapter 4 shows that the critical exchange rate is the reciprocal of the squared maximal correlation, 0.6844 and 1.4610 on the illustrative source while information rises continuously as rules split.
Data table
The numbers below are printed verbatim by the chapter's Python program and can be reproduced by running it. standard-library Python onlyprinted values that match the answer: 0.6108, 0.7234
| three-rule groupings scored | 966 |
| criterion I(T;Y) accuracy wrong features H(T) | |
| reproduce draft | 0.6108 0.780 0.3600 1.5000 |
| keep decision | 0.7234 0.805 0.4600 1.5589 |
| reproduce draft ---,-C- | --U,-CU | M--,M-U,MC-,MCU | |
| keep decision --- | --U,-C-,-CU | M--,M-U,MC-,MCU |
Show the full program
# Chapter 7. p72: Shannon (1948) set meaning aside: the engineer must
# reproduce the message, whatever it means. The bottleneck brings meaning
# back as a relevance variable Y. With the same budget of 3 rules, compare
# the grouping that best reproduces the draft itself (fewest wrong
# features M, C, U per draft, Hamming distortion) with the grouping that
# keeps most information about the gate decision. All 966 three-block
# groupings of the p11 source are scored. Illustrative.
import math
KINDS = ["---", "--U", "-C-", "-CU", "M--", "M-U", "MC-", "MCU"]
PX = [0.40, 0.18, 0.10, 0.07, 0.10, 0.06, 0.05, 0.04]
PYX = [[0.92, 0.06, 0.02], [0.25, 0.70, 0.05], [0.30, 0.55, 0.15],
[0.10, 0.60, 0.30], [0.05, 0.15, 0.80], [0.03, 0.12, 0.85],
[0.02, 0.08, 0.90], [0.01, 0.04, 0.95]]
def ham(a, b):
return len([1 for u, v in zip(KINDS[a], KINDS[b]) if u != v])
def partitions(items):
if not items:
yield []
return
first, rest = items[0], items[1:]
for part in partitions(rest):
for i in range(len(part)):
yield part[:i] + [[first] + part[i]] + part[i + 1:]
yield [[first]] + part
def scores(part):
pty = [[math.fsum(PX[x] * PYX[x][y] for x in g) for y in range(3)]
for g in part]
py = [math.fsum(r[y] for r in pty) for y in range(3)]
ity = math.fsum(v * math.log2(v / (math.fsum(r) * py[y]))
for r in pty for y, v in enumerate(r) if v > 0)
acc = math.fsum(max(r) for r in pty)
dist = math.fsum(min(math.fsum(PX[x] * ham(x, z) for x in g)
for z in range(8)) for g in part)
pt = [math.fsum(PX[x] for x in g) for g in part]
ht = 0.0 - math.fsum(p * math.log2(p) for p in pt)
return ity, acc, dist, ht
three = [p for p in partitions(list(range(8))) if len(p) == 3]
shannon = min(three, key=lambda p: (scores(p)[2], -scores(p)[3]))
relevant = max(three, key=lambda p: scores(p)[0])
print("three-rule groupings scored: %d" % len(three))
print("criterion I(T;Y) accuracy wrong features H(T)")
for name, p in [("reproduce draft", shannon), ("keep decision", relevant)]:
ity, acc, dist, ht = scores(p)
print("%-15s %6.4f %5.3f %6.4f %6.4f"
% (name, ity, acc, dist, ht))
for name, p in [("reproduce draft", shannon), ("keep decision", relevant)]:
text = " | ".join(",".join(KINDS[x] for x in sorted(g))
for g in sorted(p))
print("%-15s %s" % (name, text))Single source and citation
Fan, Y. (2026). The Information Bottleneck and Rate-Distortion Limits: How Far a Bilingual Monograph Can Be Compressed into Guardrails before Discriminative Power Falls Away (Version 1.0.0), Chapter 7: Classical Sources East and West: "Writing Does Not Exhaust Speech, Speech Does Not Exhaust Intent" and Shannon's Suspension of Meaning. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-250 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026250,
author = {Fan, Yafu},
title = {The Information Bottleneck and Rate-Distortion Limits: How Far a Bilingual Monograph Can Be Compressed into Guardrails before Discriminative Power Falls Away},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 1 · The Guardrail Source: Eight Kinds of Draft, Three Verdicts and the Counting Bound on Rules
- Chapter 2 · The Axis: Cliffs, Concave Curves and the Straight Line under Deterministic Labels
- Chapter 3 · Blahut and Arimoto: Computing the Rate-Distortion Function of a Guardrail in the Standard Library
- Chapter 4 · Phase Transitions: Maximal Correlation, the Critical Exchange Rate and Deterministic Annealing
- Chapter 5 · From Monograph to Guardrails: Agglomerative Bottlenecks, Five Knee Criteria and Bilingual Redundancy
- Chapter 6 · Opposing Evidence: Finite Samples, Learnt Compression and How Discriminative Power Is Scored
- Chapter 8 · Closing the Loop: One Compression-Recall Curve per Volume, a Knee Calculator and a Pre-Registered Protocol