What does Computing Native Chinese Copy conclude, in numbers, about “Dependency Distance: Why Europeanised Long Sentences Are Hard to Read”?
Answer: Chapter 1 shows on illustrative hand-annotated trees that Europeanised drafts have a mean dependency distance of 3.429 and native rewrites 1.515; against random orders that keep the tree fixed, the drafts score 0.950, practically random, and the rewrites 0.640.
Scope of the numbers: Its example sentences and parameters are illustrative; all calculations are generated by its programs; it is not individual legal, advertising-compliance or marketing advice, and it discusses only persuasion that keeps the rules. Chapter 2 proves that the dependency length of stacked modifiers grows roughly with the square of the layers, so that at four layers the illustrative probability of keeping every relation is only 0.098.
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: 1.515, 3.429
| sent | words | chars | DL | MDD | maxDD | de |
|---|---|---|---|---|---|---|
| A1 | 19 | 30 | 69 | 3.833 | 12 | 2 |
| A2 | 12 | 17 | 19 | 1.727 | 5 | 0 |
| B1 | 14 | 22 | 41 | 3.154 | 9 | 2 |
| B2 | 7 | 9 | 9 | 1.500 | 2 | 1 |
| C1 | 17 | 28 | 52 | 3.250 | 13 | 3 |
| C2 | 7 | 12 | 8 | 1.333 | 2 | 0 |
| D1 | 24 | 35 | 80 | 3.478 | 14 | 3 |
| D2 | 9 | 14 | 12 | 1.500 | 3 | 0 |
| mean MDD, Europeanised drafts (x1): 3.429 | ||||||
| mean MDD, native rewrites | (x2): 1.515 | |||||
| ratio rewrite / draft: 0.442 |
Show the full program
# Chapter 1. p11: Dependency distance of four Europeanised sentences and
# their native rewrites, from hand-annotated dependency trees.
# Data: illustrative sentences written and annotated by the author in a
# UD-like scheme (case markers and 的 attach to their modifier; the
# localiser heads its phrase). A token is (word, head), head 0 = root.
# Distance of an arc = |position of dependent - position of head|.
# MDD = mean dependency distance over all non-root arcs (Liu 2008).
import math
TREES = {
"A1": [("我们", 13), ("将", 13), ("为", 12), ("所有", 12), ("在", 7),
("周五", 7), ("之前", 8), ("完成", 12), ("预订", 8), ("的", 8),
("新", 12), ("会员", 13), ("提供", 0), ("一", 15), ("次", 19),
("免费", 19), ("的", 16), ("升级", 19), ("服务", 13)],
"A2": [("新", 2), ("会员", 5), ("周五", 4), ("前", 5), ("订", 0),
("好", 5), ("升级", 8), ("服务", 10), ("免费", 10), ("送", 5),
("一", 12), ("次", 10)],
"B1": [("您", 3), ("的", 1), ("订单", 12), ("已经", 12), ("被", 12),
("我们", 9), ("的", 6), ("物流", 9), ("团队", 12), ("成功", 12),
("地", 10), ("进行", 0), ("了", 12), ("处理", 12)],
"B2": [("您", 3), ("的", 1), ("订单", 5), ("已经", 5), ("寄", 0),
("出", 5), ("了", 5)],
"C1": [("通过", 14), ("使用", 1), ("我们", 6), ("的", 3), ("会员", 6),
("系统", 2), ("您", 14), ("可以", 14), ("对", 13), ("您", 13),
("的", 10), ("消费", 13), ("记录", 14), ("进行", 0),
("实时", 17), ("的", 15), ("查询", 14)],
"C2": [("会员", 2), ("系统", 3), ("里", 5), ("随时", 5), ("查", 0),
("消费", 7), ("记录", 5)],
"D1": [("如果", 15), ("您", 15), ("在", 10), ("收到", 6), ("商品", 4),
("后", 9), ("的", 6), ("七", 9), ("天", 10), ("内", 15),
("对", 14), ("商品", 14), ("的", 12), ("质量", 15), ("有", 23),
("任何", 20), ("不", 18), ("满意", 20), ("的", 18), ("地方", 15),
("您", 23), ("可以", 23), ("申请", 0), ("退款", 23)],
"D2": [("收货", 4), ("七", 3), ("天", 4), ("内", 6), ("不", 6),
("满意", 8), ("可以", 8), ("申请", 0), ("退款", 8)],
}
def distances(tree):
return [abs(i - h) for i, (_, h) in enumerate(tree, 1) if h]
def check(tree):
roots = [i for i, (_, h) in enumerate(tree, 1) if h == 0]
assert len(roots) == 1
for i in range(1, len(tree) + 1): # every token reaches the root
seen, j = set(), i
while j:
assert j not in seen
seen.add(j)
j = tree[j - 1][1]
print("sent words chars DL MDD maxDD de")
mdd = {"1": [], "2": []}
for key, tree in TREES.items():
check(tree)
d = distances(tree)
dl = int(math.fsum(d))
m = dl / len(d)
chars = int(math.fsum(len(w) for w, _ in tree))
de = len([w for w, _ in tree if w == "的"])
mdd[key[1]].append(m)
print("{:<5} {:>5} {:>6} {:>4} {:>6.3f} {:>6} {:>3}".format(
key, len(tree), chars, dl, m, max(d), de))
m1 = math.fsum(mdd["1"]) / 4
m2 = math.fsum(mdd["2"]) / 4
print("mean MDD, Europeanised drafts (x1): {:.3f}".format(m1))
print("mean MDD, native rewrites (x2): {:.3f}".format(m2))
print("ratio rewrite / draft: {:.3f}".format(m2 / m1))Single source and citation
Fan, Y. (2026). Native Cross-Context Prosody, De-Translationese Syntax-Tree Rewriting and a Conversion Corpus of Malaysian, Singaporean and Taiwanese Chinese (Version 1.0.0), Chapter 1: Dependency Distance: Why Europeanised Long Sentences Are Hard to Read. 5D Library. DOI pending
DOI pending · monograph page goes live after publication · 5DL-B-144 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026144,
author = {Fan, Yafu},
title = {Native Cross-Context Prosody, De-Translationese Syntax-Tree Rewriting and a Conversion Corpus of Malaysian, Singaporean and Taiwanese Chinese},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}Other chapter answers in this book
- Chapter 2 · Chains of De and Stacked Modifiers: an Exact Account of the Left-Branching Burden
- Chapter 3 · Fingerprints of Translationese: Ten Machine-Checkable Markers and a Poisson Likelihood-Ratio Test
- Chapter 4 · Rewriting the Tree: Three Operations and the Long-Outside Ordering Theorem
- Chapter 5 · Prosody and Breath: Segment Lengths, Dynamic Programming for Commas, and the Rhythm of Fours and Sixes
- Chapter 6 · Converting Between Malaysian, Singaporean, Taiwanese and Hong Kong Chinese: a Lexicon, an Affinity Matrix and a Library of Closing Sentences
- Chapter 7 · The Classics Examined: Sentences and Pruning in the Wenxin Diaolong, and Orwell's Six Rules
- Chapter 8 · Practice and Refutation: the Persuasive-Copy Checker, a Ceiling on Private Follow-Ups, and a Pre-Registered Test