Series ID: EDKS-ADV-VOC-ZH-0027 · Advanced English Vocabulary (Chinese Edition) · Lesson No.027 · Batch 07 · C1 → C2 · 简体中文辅助
Knowing what a word means is not enough. Advanced control also means knowing what kinds of situations that word normally lives in.
知道一个词“是什么意思”还不够。高级词汇控制还包括：知道这个词平时都跟什么样的情境、评价与话语环境一起出现。
You look up a word. The dictionary definition seems correct. The grammar is correct. The collocation is possible. Yet your sentence still feels slightly wrong.
Why?
One reason may be that the word usually appears in a different semantic environment.
Consider the expression bent on. Oxford’s explanation of the British National Corpus uses concordance evidence to show that bent on is often followed by actions with negative implications, such as destruction or other harmful intentions. A learner who treats bent on as simply “determined to” may therefore produce a grammatically correct but pragmatically odd sentence such as She was bent on helping every child succeed. The meaning is understandable, but the phrase carries a usage history that makes the sentence sound unusual.
这种“一个词经常和什么样的语义环境一起出现”的现象，就是本课要研究的对象之一。
Corpus linguists use related terms such as semantic preference and semantic prosody. These terms are not perfectly identical, and modern research continues to debate how semantic prosody should be defined and measured. A 2025 paper in Research Methods in Applied Linguistics explicitly notes that semantic-prosody categorisation has methodological reliability issues and that there is no complete consensus on operationalisation. A 2024 Journal of Pragmatics paper likewise argues that semantic prosody is theoretically more complex than a simple positive/negative label.
Do not memorise “this word is negative”. Learn what kinds of meanings, situations and evaluations repeatedly surround it.
不要死记“这个词是负面的”。要看它反复跟什么样的意思、场景与评价一起出现。
This lesson owns the Mandarin-supported C1–C2 route for semantic prosody, semantic preference and typical-context control. It does not replace Lesson No.006 on near-synonyms, Lesson No.009 on corpus evidence, Lesson No.012 on idioms and figurative meaning, Lesson No.013 on polysemy, or the existing owner for connotation. Here the narrower question is: what does repeated usage around a word teach us that a dictionary definition alone may not?
1. Start with collocation | 从搭配开始
Oxford defines a collocation as a combination of words that occurs frequently and more often than would be expected by chance. Examples include combinations such as resounding success or crying shame.
Collocation answers:
Which words tend to occur together?
2. Semantic preference asks a wider question | semantic preference 问的不只是“哪个词”
Instead of looking only at individual collocates, semantic preference looks for the semantic class of the company a word keeps.
For example, a word may repeatedly occur around:
- problems;
- risks;
- physical movement;
- evaluation;
- measurement;
- institutional processes;
- human emotion.
3. Semantic prosody adds evaluative/discourse tendency | semantic prosody 再往前一步：这些环境是否形成某种评价或话语倾向？
A seemingly neutral expression may repeatedly occur in contexts associated with difficulty, harm, success, benefit, uncertainty or another evaluative orientation.
But treat this as a tendency supported by usage evidence, not a fixed emotional tag.
4. Connotation is related but not identical | connotation 与 semantic prosody 有重叠，但不是同一个概念
Connotation concerns associations and evaluative meanings attached to a word. Semantic prosody is typically investigated through repeated usage patterns across corpus contexts.
A word may have an obvious negative connotation without needing corpus evidence. Semantic-prosody analysis becomes especially useful when the word looks relatively neutral in isolation but develops a patterned evaluative environment in use.
5. Do not collapse everything into “positive / negative” | 正面/负面只是最粗的一层
More useful categories may include:
- problem-oriented;
- risk-oriented;
- benefit-oriented;
- institutional;
- cautious;
- critical;
- neutral-technical;
- conflict-related;
- unexpected/marked;
- human-agentive;
- process-oriented.
6. The same word can have different prosodies by sense | 多义词不同义项可以有不同“典型环境”
Take commit:
- commit a crime/error — one sense pattern;
- commit resources — another;
- commit to a plan — another grammatical and semantic pattern.
Never assign one global prosody to a polysemous word without separating senses.
7. Grammar can separate prosodies | 结构不同，语义环境也可能不同
Compare:
- subject to approval;
- subject somebody to pressure;
- the subject of the article.
Same spelling, different structure, different semantic environment.
8. Semantic prosody is often sense + frame specific | 最稳的学习单位是“义项＋结构＋典型环境”
word + sense + frame + typical context
单词＋义项＋结构＋典型语境
9. Oxford’s BNC example: bent on | 一个很好的真实例子：bent on
Oxford explains that concordance evidence for bent on shows many following actions with negative implications. This tells the learner something beyond a simple translation such as “determined to”.
Natural learner conclusion:
bent on often fits strong determination in problematic or disapproved contexts.
Important word: often, not always.
10. Why this matters for production | 为什么 C1–C2 特别需要这层知识？
Because advanced learners increasingly produce sentences that are grammatically correct but distributionally unusual.
The sentence can pass grammar checking and still sound odd because the lexical choice belongs to another discourse environment.
11. “Correct meaning” can still be wrong usage | 字典义对了，语境仍然可能错
Suppose a learner selects a verb whose definition matches “make worse”. If the verb is strongly conventional in formal problem-oriented contexts, it may sound unnatural in a casual positive anecdote.
Meaning compatibility is necessary but not sufficient.
12. Typical context is part of lexical depth | 词汇深度应该包含“这个词通常在哪里出现”
- definition;
- sense;
- collocation;
- grammar frame;
- register;
- domain;
- typical semantic environment;
- evaluation/discourse tendency.
13. Use corpus lines, not intuition alone | semantic prosody 特别需要多条语料证据
Search a corpus or concordance and read multiple lines. Ask:
- What semantic classes recur?
- What evaluation surrounds the word?
- What grammatical frame is stable?
- Which patterns are genre-specific?
- Are there important counterexamples?
14. Do not decide from five convenient examples | 五条例句不能证明一个词“天生负面”
Corpus sampling matters. The 2025 methodological study on semantic prosody highlights reliability concerns in categorisation. Advanced learners should therefore treat small samples as exploratory evidence.
15. Genre can change the pattern | 新闻、学术、小说、会议里的环境可能不同
A word that appears around risk in financial reporting may occur differently in fiction or everyday conversation.
Always ask: Which corpus? Which genre? Which sense?
16. Domain can change the pattern | 专业领域会重写普通词的语义邻居
Exposure in photography, finance, medicine and general English occurs around very different concepts.
Lesson No.019’s domain labels apply directly.
17. Register can change the pattern | 同一个词在口语与正式写作里的“常见邻居”也会不同
Do not promote a phrase into conversational English just because it is frequent in formal institutional prose.
18. Time can change the pattern | 语义倾向也可能随时代变化
Corpora represent periods of language use. A historical tendency may weaken, strengthen or shift. Prefer current corpora when your goal is contemporary production.
19. Variety can change the pattern | 英式、美式、全球英语也可能有差异
Do not turn one national corpus pattern into a universal English rule without checking.
20. Frequency does not equal prohibition | 典型环境不是禁令
A phrase can appear outside its usual environment for irony, contrast, creativity or deliberate markedness.
The advanced question is not “Is this allowed?” It is:
“If I use it here, will the reader hear an unusual tone—and do I want that?”
21. Marked usage can be powerful | 违反典型语境，有时就是修辞
A writer may deliberately place a normally problem-oriented expression in a positive context for humour or irony.
To break a pattern effectively, you first need to know the pattern exists.
22. Semantic prosody is expectation management | 学会以后，你会开始预测“这个词后面通常会出现什么类型的信息”
That prediction helps both reading and listening.
23. Typical contexts speed comprehension | 常见语义环境会减少实时解释成本
If a phrase usually appears around risk, encountering it activates a useful expectation. The listener can allocate attention to the specific risk rather than rebuilding the whole frame.
24. Unexpected context creates a signal | 当词出现在反常语境，读者会注意到
This may indicate:
- irony;
- humour;
- creative metaphor;
- semantic change;
- unusual individual style;
- learner error.
25. Learn “typical neighbours” | 一个高级词条应该记录 3–5 个典型邻居
Example structure:
| Target | Typical semantic neighbours |
|---|---|
| mitigate | risk, impact, harm, effects, consequences |
| exacerbate | problem, tension, inequality, symptoms, risk |
| plausible | explanation, account, interpretation, hypothesis |
| robust | method, evidence, system, result, framework depending on domain |
These are teaching examples, not exhaustive corpus claims. Verify in your target register.
26. Learn semantic class, not only list | 比五个搭配更重要的是看见它们属于哪一类
Mitigate often relates to reducing undesirable effects. Plausible often evaluates explanations, accounts or interpretations. The category is more transferable than the memorised list.
27. Contrast semantic environments | 把近义词放到不同语义环境里对比
Possible pair:
- mitigate risk;
- reduce cost.
Both may mean “make less”, but their typical objects and register differ.
28. Near-synonyms can differ mainly in environment | 有时两个词定义很近，真正差别在“常跟谁一起出现”
This is why a thesaurus is dangerous at C1–C2. It can identify semantic neighbours without showing distributional fit.
29. Corpus evidence can reveal hidden asymmetry | 语料能告诉你：A 可以换 B，不代表 B 常换 A
Two words may overlap in one narrow context but have very different wider distributions.
30. Do not infer prosody from dictionary synonyms | 同义词栏不能直接告诉你 usage prosody
Use corpus examples and learner-dictionary usage notes after synonym discovery.
31. Use a 20-line concordance scan | 20 条语料线是一个实用起点
This is a learner heuristic, not a methodological standard.
- Read 20 randomly selected lines.
- Group collocates by semantic class.
- Mark evaluation if visible.
- Separate senses/frames.
- Look for counterexamples.
32. Then expand the sample if the pattern matters | 高风险词不要停在二十条
If the word affects academic stance, professional reputation or exam precision, inspect more lines and more than one corpus/source.
33. Sort by sense first | 语料分析第一步：不要把不同义项混在一起
Polysemy can create a false pattern if multiple senses are mixed.
34. Sort by grammar frame next | 第二步：分结构
Verb + object, adjective + noun, noun + preposition and phrasal constructions may behave differently.
35. Sort by register/domain next | 第三步：分语域/领域
Academic, journalistic and conversational lines should not always be merged.
36. Only then infer a tendency | 最后才说“这个词似乎通常…”
Use cautious language:
- often occurs with…;
- commonly appears in…;
- tends to be associated with…;
- in this corpus/register…
37. Avoid “always negative” claims | 世界级标准：除非证据真的支持，否则不要绝对化
The 2025 reliability paper is a useful reminder that prosody categorisation itself can be difficult. Learners should model epistemic caution.
38. Learn semantic prosody with examples and non-examples | 典型例子与反例一起存
Target: bent on.
- Typical-style example: bent on revenge/destruction.
- Marked/non-typical example: bent on kindness.
Ask how the second sentence feels and why.
39. Marked examples teach the edge of the pattern | 反例不是为了证明规律错，而是为了看规律边界
40. Build “neutral definition + typical environment” cards | 卡片不要只写中文翻译
| Field | Example |
|---|---|
| Target | bent on |
| Core meaning | strongly determined to do something |
| Typical environment | often negative/problematic intentions in corpus evidence |
| Register | neutral/formal-ish phrase depending on context |
| Warning | positive uses may sound marked |
| My test sentence | He seemed bent on escalating the dispute. |
41. Add confidence labels to prosody claims | 语义倾向笔记要标证据强度
- Strong pattern;
- Moderate pattern;
- Possible tendency;
- Not enough evidence.
42. Add corpus/source label | 写清楚这个结论来自哪里
Example:
BNC examples / news corpus / academic corpus / my field corpus.
43. Compare two corpora when possible | 不同 corpus 的差异本身就是学习
If a pattern appears in news but not academic prose, you have learned a register distinction.
44. Compare L1 intuition with L2 evidence | 中文直觉经常会把两个英文词压成同一个翻译
Suppose two English words both translate roughly as “严重”. Their typical nouns may differ dramatically. Corpus evidence can expose the difference.
45. Translation equivalence can hide prosody mismatch | “中文意思一样”不代表语境倾向一样
Always check the English distribution of the candidate word.
46. Semantic prosody matters in translation | 翻译时需要保留的不只是字典义，还有语境气质
A translator choosing a semantically correct word with the wrong evaluative distribution may subtly change the author’s stance.
47. Semantic prosody matters in paraphrase | 第017课改述时，也必须保护典型语境与评价
Replacing a neutral word with one that regularly appears in stronger negative environments can intensify the passage without changing the dictionary core.
48. Semantic prosody matters in stance | 第020课的立场控制也会被“词的典型邻居”影响
A word that often appears in critical discourse may make your sentence sound more evaluative than you intended.
49. Semantic prosody matters in professional tone | 专业场合里，词选错会显得过度负面或过度积极
Risk reports, performance reviews and feedback documents are especially sensitive to evaluative lexical patterns.
50. Semantic prosody matters in diplomacy | 会议里“ technically correct ”的词也可能太强
If a word regularly appears in conflict-oriented discourse, a milder alternative may better preserve collaboration.
51. Semantic prosody matters in academic criticism | 学术批评需要知道“质疑、挑战、批评”的常见语境差异
Corpus evidence can help distinguish the typical environments of verbs such as question, challenge, criticise.
52. Semantic prosody matters in learner writing | 很多“不自然”的高级作文，其实不是语法错
The learner selected a plausible dictionary synonym that lives in the wrong discourse neighbourhood.
53. Build prosody-aware synonym sets | 近义词表里加“典型环境”一栏
| Word | Core distinction | Typical environment |
|---|---|---|
| problem | general difficulty | broad |
| constraint | factor limiting options | planning, technical, analytical |
| bottleneck | limiting point in a process | systems/process/performance |
| crisis | severe unstable situation | high-stakes negative |
The table shows a teaching model, not a complete corpus profile. Verify in your target domain.
54. Learn semantic preference through categories | category-based notes are more transferable
Instead of memorising:
mitigate risk, mitigate harm, mitigate impact, mitigate effects
learn:
mitigate typically combines with undesirable outcomes or conditions that can be reduced in severity.
55. Then generate new collocations cautiously | 有了 semantic class 后，可以预测新搭配，但必须验证
You may predict mitigate the consequences. Then verify it.
56. Prediction is useful because it turns vocabulary into a model | 你开始理解“为什么这个词会跟这些对象一起出现”
57. Use semantic-prosody mismatches as editing alarms | 写作时出现“意思对，但就是怪”的感觉，值得查 corpus
Especially check when:
- you used a thesaurus;
- you translated directly from Chinese;
- you chose a rare formal word;
- you are writing evaluative feedback;
- you changed a near-synonym;
- the sentence is grammatically correct but sounds off.
58. Do not search the whole web blindly | 普通搜索结果会被 SEO、新闻热点和重复内容扭曲
Prefer balanced corpora, learner-dictionary corpora, domain corpora or corpus tools that let you inspect concordance lines.
59. Use SKELL or Sketch Engine-style evidence when available | 语料工具帮助你看真实搭配分布
Search target words, inspect word sketches/collocates and read actual concordance lines before deciding what the pattern means.
60. Frequency lists are only the first layer | 词频告诉你“常不常见”，prosody 告诉你“常见在哪里”
61. Oxford 3000/5000 help prioritise words, not prosody alone | 词表与语义倾向承担不同任务
Oxford’s advanced word lists use corpus frequency and learner relevance to prioritise vocabulary. For prosody, you still need contextual evidence.
62. Oxford Phrase List reminds us to learn beyond single words | common phrases are part of the lexical unit
The Oxford Phrase List includes collocations, idioms, phrasal verbs and common prepositional phrases. Prosody often belongs to the phrase, not the isolated headword.
63. Learn phrase-level prosody | 语义倾向经常属于“整块表达”
- bent on;
- give rise to;
- subject to;
- pose a threat;
- cast doubt on.
Analyse the phrase as the unit when evidence supports that approach.
64. Prosody can belong to a grammatical construction | 有时“词＋语法框架”共同产生语境倾向
Do not detach the word from its construction during analysis.
65. Build a semantic-environment map | 建立“语义邻里图”
Target: mitigate.
- undesirable outcome;
- risk;
- harm/impact;
- severity reduction;
- professional/technical register.
Now compare with eliminate, whose meaning implies removal rather than reduction.
66. Build a prosody contrast map | 两个近义词并排看“常见世界”
Target pair: address / solve.
- address → deal with/take action on problem/issue/concern;
- solve → reach successful resolution of problem/question/puzzle.
These typical environments reinforce the semantic boundary already learned elsewhere.
67. Build a phrase-completion prediction test | 看前半段，预测后面属于什么语义类别
- mitigate the ___ → likely undesirable condition/outcome;
- plausible ___ → explanation/account/interpretation;
- pose a ___ → often challenge/risk/threat/problem depending on context;
- strongly ___ → verb/adjective patterns vary by register.
Then verify, do not assume.
68. Use prosody cloze exercises | 语义类别填空比单词翻译更接近真实选择
Prompt:
The new safeguards may ______ the risk of accidental disclosure.
Candidate analysis:
- mitigate — semantically and distributionally plausible;
- celebrate — grammatical form possible? no semantic fit;
- eliminate — stronger claim;
- address — different relation.
69. Use corpus-before-production drills | 新高级词先观察 10–20 条，再自己写
This reduces the chance of producing a dictionary-correct but usage-odd sentence.
70. Then use production-before-corpus sometimes | 反过来也要测自己的直觉
- Write three sentences.
- Predict which one is most typical.
- Check corpus evidence.
- Explain the mismatch.
71. This trains distributional intuition | 最终目标不是永远查 corpus，而是逐步内化常见模式
72. Use reading harvest for prosody | 第022课阅读系统增加一个 R+ 环境标记
When you meet a useful word, ask not only “what does it mean?” but “what kind of paragraph is it living in?”
73. Use listening harvest for prosody | 第023课也一样：语气与语境一起听
Does the phrase appear during disagreement, repair, explanation, risk discussion or praise?
74. Use speaking to test whether your prosody knowledge is active | 口语里最容易暴露“这个词会不会跑错场”
After speaking, log cases where listeners reacted oddly or you immediately wanted to replace a word.
75. Use writing revision to catch prosody mismatch | 写作慢下来后做“典型语境审计”
- Did I choose this word from a thesaurus?
- Does the noun/verb partner belong to its normal semantic class?
- Does the phrase create a stronger evaluation?
- Is the register appropriate?
76. Worked case: Maren writes “bent on helping” | 案例一
Maren finds bent on translated as “determined to”. She writes:
She was bent on helping every student improve.
The sentence is understandable. Corpus evidence suggests the phrase often appears with negative/problematic intentions, so the sentence may sound marked.
Natural alternative:
She was determined to help every student improve.
中文分析：词典义没有错。真正的问题是典型语境不同。
77. Worked case: Iona treats prosody as a fixed rule | 案例二
Iona learns “bent on = negative” and begins marking every positive corpus example as “wrong English”.
Her tutor repairs the concept:
- prosody = tendency;
- corpus = evidence;
- marked use can be deliberate;
- sense/register/genre matter.
She replaces absolute notes with probabilistic notes.
78. Worked case: Leonie uses a critical word in feedback | 案例三
Leonie chooses a near-synonym that repeatedly appears in strongly critical contexts. The feedback becomes harsher than intended.
She checks alternative words whose typical environments are more neutral and selects one that matches the intended interpersonal stance.
79. Worked case: domain shift changes the neighbourhood | 案例四
A learner studies exposure in general English, then enters finance and sees market exposure, currency exposure, risk exposure. The word’s domain-specific semantic preference becomes part of specialist vocabulary.
80. The FENCE semantic-prosody system | FENCE 语义倾向系统
| Fence | Question | Action |
|---|---|---|
| F0 Sense | Which meaning is active? | separate senses |
| F1 Frame | Which grammatical construction? | separate structures |
| F2 Collocation | Which words recur nearby? | collect neighbours |
| F3 Preference | What semantic classes recur? | group collocates |
| F4 Prosody | What evaluative/discourse tendency appears? | label cautiously |
| F5 Register | Which genre/domain/variety? | context label |
| F6 Counterexample | What breaks the pattern? | inspect marked uses |
| F7 Verification | Is the sample large/diverse enough? | expand corpus evidence |
| F8 Transfer | Can I use the word naturally in a new context? | production test |
81. Build a semantic-prosody card | 建立“语义邻里卡”
| Field | Example |
|---|---|
| Target | bent on |
| Sense | strongly determined |
| Frame | bent on + noun/-ing |
| Typical semantic class | strong intentions/actions |
| Observed tendency | often negative/problematic in BNC evidence |
| Counterexample | positive uses possible but may sound marked |
| Source | Oxford BNC explanation |
| Confidence | moderate/strong depending on corpus sample |
82. A 15-minute prosody investigation | 15 分钟语义倾向训练
| Time | Task |
|---|---|
| 0–3 min | Choose one target word/phrase. |
| 3–7 min | Read 15–20 concordance lines. |
| 7–10 min | Group collocates by semantic class. |
| 10–12 min | Identify possible evaluative tendency. |
| 12–15 min | Write one typical and one marked sentence. |
83. A 30-minute corpus comparison workshop | 30 分钟语料对比
| Time | Task |
|---|---|
| 0–5 min | Choose a near-synonym pair. |
| 5–12 min | Inspect 20 lines for Word A. |
| 12–19 min | Inspect 20 lines for Word B. |
| 19–24 min | Compare semantic classes and register. |
| 24–27 min | Write contrast rule. |
| 27–30 min | Produce 2 new sentences each. |
84. A seven-day semantic-environment challenge | 七天语义环境挑战
| Day | Focus |
|---|---|
| 1 | collocation vs semantic preference |
| 2 | connotation vs semantic prosody |
| 3 | polysemy + frame separation |
| 4 | near-synonym corpus comparison |
| 5 | register/domain variation |
| 6 | marked counterexamples |
| 7 | independent prosody audit of own writing |
85. A 30-day semantic-prosody challenge | 30 天语义倾向升级
- Investigate 20 high-value words/phrases.
- Record 5–10 typical collocates for each.
- Group each item’s collocates into semantic classes.
- Compare 10 near-synonym pairs.
- Find at least one counterexample for every strong tendency claim.
- Label register/domain for each entry.
- Correct 10 thesaurus-driven sentences using corpus evidence.
- Build 10 meaning+frame+context cards.
86. A 12-week progression | 12 周语义倾向路线
| Weeks | Focus | Evidence of progress |
|---|---|---|
| 1–2 | collocation and semantic classes | less word-by-word synonym choice |
| 3–4 | prosody/connotation distinction | more nuanced evaluation |
| 5–6 | sense/frame separation | fewer false global rules |
| 7–8 | genre/domain comparison | better register control |
| 9–10 | production and markedness | more natural advanced word choice |
| 11–12 | independent corpus inference | learner tests intuition rather than guessing |
87. The first weak link diagnostic | 第一个弱点诊断
| Symptom | Likely weak link | Repair |
|---|---|---|
| My sentence is grammatical but sounds odd. | typical-context mismatch | corpus scan |
| I label words simply positive/negative. | prosody oversimplification | semantic-class analysis |
| I use one prosody for every sense. | polysemy | split sense/frame |
| I trust a thesaurus synonym. | distribution ignorance | compare collocates |
| My formal writing sounds too harsh. | evaluative environment | register/prosody comparison |
| I generalise from five examples. | sample weakness | expand corpus sample |
| I think unusual usage is “wrong”. | markedness misunderstanding | counterexample analysis |
| My corpus result changes by source. | genre/domain variation | label corpus and register |
88. Practice A — identify the lexical layer | 练习 A：这是 collocation、semantic preference 还是 prosody？
- compelling evidence occurs frequently.
- A verb repeatedly takes nouns meaning problems or risks.
- An apparently neutral phrase often occurs in disapproving discourse.
- A word is associated with sadness because that is part of its conventional connotation.
89. Practice B — separate sense/frame | 练习 B：不要混语义环境
- commit a crime / commit resources / commit to a plan;
- subject to approval / subject someone to pressure / subject of a report;
- address a concern / address an audience / address on an envelope.
For each, create separate sense+frame entries.
90. Practice C — typical or marked? | 练习 C：典型还是刻意反常？
- He was bent on revenge.
- She was bent on making everyone feel welcome.
- The policy mitigates the risk of delay.
- The celebration mitigates the joy of the event.
Discuss distributional naturalness, not only literal grammaticality.
91. Practice D — corpus claim language | 练习 D：用科学谨慎语言描述趋势
- Rewrite “X is always negative.”
- Rewrite “Y only occurs in academic writing.”
- Rewrite “Z can never be used positively.”
92. Practice E — prosody-aware synonym choice | 练习 E：同义替换前检查语义环境
- problem / constraint / crisis;
- reduce / mitigate / eliminate;
- possible / plausible / likely;
- question / challenge / criticise.
For each set, write one sentence where each word is the best choice and explain the typical context.
93. Answers A | 练习 A 参考
- Collocation.
- Semantic preference.
- Semantic prosody / evaluative discourse tendency, if supported by sufficient corpus evidence.
- Connotation.
94. Answers D — model cautious claims | 练习 D 参考
- “In this corpus and sense, X often occurs in negatively evaluated contexts.”
- “Y is especially frequent in academic writing in the corpus examined, though it also occurs elsewhere.”
- “Positive uses of Z appear less typical in the sample and may sound marked, but they are possible.”
95. How a strong tutor teaches semantic prosody | 高水平导师怎样教“词的典型语境”？
A weak lesson gives a list: “positive words / negative words”. A stronger lesson asks:
- Which sense?
- Which grammatical frame?
- Which corpus?
- What collocates recur?
- What semantic class do they form?
- What evaluative tendency appears?
- What counterexamples exist?
- Is the pattern stable across genres?
The tutor teaches evidence-based distributional judgement.
96. The learner should become their own distribution analyst | 最终，学习者要会问“这个词通常生活在哪个世界？”
Before adopting a new advanced word, ask:
- Which sense am I using?
- What grammar frame?
- Which nouns/verbs surround it?
- What semantic class do they belong to?
- What tone or evaluation tends to emerge?
- Does my target genre match that pattern?
- Would my sentence sound marked?
- If marked, is that intentional?
97. Canonical ownership and collision boundary | 本课的所有权边界
This lesson owns the Mandarin-supported C1–C2 route for semantic prosody, semantic preference and typical-context control. It coordinates but does not replace:
- Lesson No.006 — near-synonym precision.
- Lesson No.009 — corpus verification.
- How Connotation Changes Meaning Beyond Dictionary Definitions — connotation mechanism.
- Lesson No.013 — polysemy and familiar-word new senses.
- Lesson No.017 — meaning-preserving paraphrase.
- Lesson No.020 — stance and evaluation vocabulary.
98. Recommended reference floor | 推荐参考资源
- Oxford Learner’s Dictionaries — The British National Corpus: facts and figures — explains corpus use for meaning, grammar and collocation and gives the bent on example with a recurrent negative semantic environment.
- Oxford Advanced Learner’s Dictionary — collocation — definition of recurrent word combinations and their importance for advanced learners.
- Oxford Phrase List — common collocations, phrases, idioms, phrasal verbs and prepositional phrases aligned to CEFR.
- Russnes (2025) — Semantic prosody, categorisation and inter-rater reliability — methodological cautions around categorising semantic prosody.
- Grabowski & Trklja (2024) — Unravelling the complexity of semantic prosody — theoretical discussion showing that semantic prosody is more complex than a simple evaluative label.
- Acquisition of conscious and unconscious knowledge of semantic prosody — experimental evidence that repeated linguistic contexts can contribute to learned emotional/evaluative associations.
- Oxford Advanced Learner’s Dictionary — sense, labels, examples and collocation verification.
- SKELL / Sketch Engine — corpus-based examples and collocational evidence for learner investigation.
99. SEO language map | 本课关键词范围
This lesson naturally serves semantic prosody English, semantic preference English, collocation and semantic prosody, corpus vocabulary learning, advanced word choice corpus, typical word contexts, word tone English, positive negative semantic prosody, collocation C1 C2, semantic prosody for Chinese learners, advanced vocabulary naturalness, corpus-based synonym choice, thesaurus mistakes advanced English, typical collocates English, semantic environment vocabulary, 高级英语词汇, 语义倾向, 语义韵, 语义偏好, 英语搭配, 语料库选词, 英语词语气质, 英语典型语境, 中文母语英语学习, C1 英语 and C2 英语.
100. Your assignment | 本课作业
Choose three high-value words or phrases that you understand but do not fully trust in production.
- Write the core dictionary meaning.
- Separate the relevant sense and grammar frame.
- Collect 20 corpus/concordance examples for each item.
- Group nearby words into semantic classes.
- Note any recurrent evaluative/discourse tendency.
- Find at least two counterexamples.
- Label register/domain/source.
- Write one cautious tendency statement using often/tends to/in this corpus.
- Write one typical sentence.
- Write one deliberately marked sentence.
- Explain how the marked sentence changes tone.
- Compare one near-synonym using the same process.
Then return to one piece of your own writing and identify three thesaurus-driven or translated word choices. Check whether the words live in the semantic environments you intended.
你的目标不是给每个词贴上“正面”“负面”标签。你的目标是逐渐建立一种分布直觉：这个词通常跟谁一起出现、在哪种话语里出现、出现以后读者会预期什么。
101. The final rule | 最后一条规则
Learn the word. Then learn the world around the word.
先学这个词，再学这个词周围的世界。
Dictionary meaning tells you what a word can mean. Grammar tells you how it can be built into a sentence. Collocation tells you which words it often combines with. Semantic preference tells you which semantic classes repeatedly surround it. Semantic prosody asks whether those repeated environments create a broader evaluative or discourse tendency.
None of these should become a rigid rule.
But together they explain why advanced English is not simply a larger vocabulary list.
It is a growing model of where words belong.
真正的 C1–C2 词汇控制，不只是“这个词我认识”。而是“我知道这个词通常出现在哪里，和什么样的意思在一起，什么时候自然，什么时候会显得刻意”。
Series: EDKS-ADV-VOC-ZH · Lesson 027 · Batch 07.
