Series ID: EDKS-ADV-ZH-0006 · How to Learn Advanced English (Chinese Edition) · Lesson No.006 · C1 → C2

Speak Without Translating in Real Time | 实时英语表达：不用先翻译再开口

Fluent speaking does not mean speaking without thinking. It means thinking and speaking can run together without every idea first becoming a complete Chinese sentence.

流利口语，不是“完全不思考”。真正的流利，是思考和表达能够同时运行，而不是每一个想法都必须先变成完整中文句子，再翻译成英语。

Many advanced Mandarin-speaking learners know enough English to speak well. The problem is timing. The idea appears in Chinese, a complete Chinese sentence forms, the learner searches for English vocabulary, changes the grammar, checks whether it sounds natural, and only then begins to speak. By that time, the conversation has moved.

很多中文母语高级学习者其实“英语够用了”，但实时交流还是慢。原因不是没有知识，而是路线太长：先出现中文意思 → 形成完整中文句子 → 找英文词 → 改英语语法 → 检查自然度 → 最后才开口。等你准备好，对话已经走到下一步。

This lesson shortens the route.

本课要做的不是“禁止中文”，而是把实时表达改造成：

idea → English meaning chunks → sentence growth → interaction → repair.

想法 → 英语意义词块 → 句子逐步长出来 → 互动 → 自我修正。

You will still use Mandarin when it helps you think deeply. But in real-time speaking, you will increasingly stop waiting for a complete Chinese sentence before English begins.

Before we begin: this lesson owns real-time production | 开始之前：本课负责“实时口语生成层”

eduKate Sengkang already owns school-stage oral communication, planned response, spoken interaction, academic discussion and general oral-English mechanisms. This page does not replace them.

eduKate Sengkang 已经有学校阶段 oral communication、planned response、spoken interaction、academic discussion 以及一般口语机制页。本课不复制这些内容。

This lesson owns one advanced Mandarin-speaker problem:

How do I convert ideas into English fast enough to speak, interact, repair and negotiate meaning without first constructing a full Chinese sentence?

怎样让想法直接进入英语，使我能及时说、及时回应、及时修正和协商意义，而不是先完整造中文再翻译？

How Oral English Builds Thinking and Expression — general oral-language owner.
Secondary 4 English Learning Guide | Oral Communication: Planned Response, Spoken Interaction and Recovery — school-stage owner.
Secondary 3 English Learning Guide | Academic Discussion and Seminar — stage-specific academic discussion owner.
Lesson No.005 — advanced listening, stance and fast speech.
Lesson No.003 — Mandarin→English sentence reconstruction.

1. C1–C2 speaking is interaction, not a monologue contest | C1–C2 口语不是“连续说很久比赛”

Cambridge C1 Advanced Speaking requires several different speaking modes: answering questions, giving a long turn, comparing and speculating, interacting with another candidate, negotiating a decision and discussing an issue further.

剑桥 C1 Advanced 口语不是只看“能不能说一分钟”。它要求你回答、比较、推测、表达观点、回应别人、协商、评价、共同做决定，并进一步讨论。

The CEFR C1 spoken descriptors similarly emphasise range, accuracy, fluency, interaction and coherence. At C2, reformulation becomes especially important: highly proficient speakers can backtrack around difficulty and express fine shades of meaning without visibly collapsing the conversation.

CEFR C1 也把口语能力拆成 range、accuracy、fluency、interaction、coherence。到了 C2，一个核心能力是 reformulation：遇到表达困难时，能绕开、重说、换一种形式继续，而不是整段崩掉。

So advanced speaking is not “perfect English spoken fast”. It is meaning kept alive under time pressure.

高级口语不是“完美英语 + 很快”。真正的高级，是在时间压力下仍然让意义继续活着。

2. Fluency has several parts | 流利度不是只有“语速”

Second-language fluency research often distinguishes dimensions such as speed, breakdown and repair.

Dimension	Question	中文版理解
Speed fluency	How efficiently does speech move?	整体速度是否顺畅？
Breakdown fluency	Where and how often do pauses interrupt speech?	在哪里停？停多久？
Repair fluency	How do repetitions, restarts and corrections behave?	改口、重复和修正怎样发生？

A slower speaker can be highly fluent if pauses are well placed and the message remains coherent. A fast speaker can sound disfluent if speech repeatedly collapses, restarts or becomes hard to follow.

所以“每分钟多少词”不能单独代表高级口语。停顿位置、修复方式和意义连续性同样重要。

3. Chinese learners themselves report the translation bottleneck | 中国学习者自己也报告“先中文再翻译”瓶颈

A national mixed-methods study of 1,525 Chinese college students found that speaking was a major challenge for many participants. Among reported speaking difficulties were limited speaking environments, limited practice and L1 interference. Some participants explicitly described forming the idea in Chinese first, translating it into English and only then speaking.

一项全国性中国大学生混合研究调查了 1,525 名学习者。很多人把口语列为主要困难之一，并报告了缺少英语环境、口语练习不足以及 L1 interference。部分学习者非常明确地描述了这条路线：先在脑中形成中文 → 翻译成英语 → 最后才说出来。

That does not mean every Chinese learner follows this route. But if you recognise yourself in it, the process can be retrained.

这不代表每个中国学习者都一样。但如果你发现自己正在使用这条路线，它是可以重新训练的。

4. Do not ban Chinese thinking | 不要禁止中文思考

Your strongest conceptual language may still be Mandarin in some domains. That is not a defect.

复杂哲学、专业知识、个人记忆或文化经验，有时中文仍然是你更强的思考语言。这不是问题。

The real-time speaking problem begins when you insist on creating a complete Chinese sentence before English production can start.

真正需要改变的是：

不要等完整中文句子完成以后，英语才开始。

5. Separate concept from sentence | 把“概念”与“句子”分开

Before speaking, you do not need a complete sentence in any language. You need a conceptual packet.

For example:

idea: policy useful / implementation expensive / support with conditions

Then English can grow:

I think the policy is useful, but I’d support it only if the implementation cost can be controlled.

没有必要先在脑中完整生成：

我认为这个政策有用，但是只有在实施成本能够得到控制的情况下我才会支持它。

高级口语训练要逐步从“句子翻译”转向“概念直接编码”。

6. Build English from nodes | 用英语节点建句

Use speaking nodes:

position
reason
example
limit
next step

Example topic: remote work.

Nodes:

useful / autonomy / not every role / hybrid / evidence by task

Spoken answer:

I think remote work can be useful because it gives people more autonomy. But it doesn’t fit every role, so I’d probably favour a hybrid model and judge it by the type of work being done.

先有节点，再让句子长出来。

7. Speak before the sentence is complete | 不要等整句完美以后才开口

Native and highly proficient speakers plan incrementally. They often begin a clause while later material is still being planned.

Advanced learners should practise the same principle:

start with a safe launch phrase, then build forward.

Examples:

I think the main issue is…
The way I see it…
There are two reasons for that…
What matters most here is…

这些不是为了背模板，而是让你获得“先开口，再继续规划”的时间。

8. Launch phrases reduce planning pressure | 起步词块可以降低实时规划压力

A useful launch phrase does three things:

signals your discourse function
buys a small amount of planning time
creates a grammatical route for what follows

例如 One reason I say that is… 已经告诉听者：接下来是 reason，而且帮你建立从句框架。

9. Lexical bundles can support oral fluency | 词块可以支持口语流利度

Research on L2 oral fluency has found meaningful relationships between multiword lexical bundles and aspects of speed, breakdown and repair fluency.

2024 年 System 的研究进一步显示，多词词块与 L2 口语流利度多个维度存在关系。对高级学习者的启示很实际：不是所有语言都应该每次现场从零拼装。

Useful bundles include:

as far as I can tell
the main point is
what I mean is
it depends on whether
one way of looking at it is
the difficulty with that is
from a practical point of view

词块是实时表达的“预编译单位”。

10. But do not turn chunks into scripts | 词块不是固定脚本

If every answer begins with the same five memorised expressions, fluency becomes formulaic.

The goal is a flexible inventory:

multiple ways to agree
multiple ways to disagree
multiple ways to qualify
multiple ways to reformulate

高级口语需要选择，不是重复。

11. Build direct concept-to-English routes | 建立“概念 → 英语”直接路线

For high-frequency concepts, stop requiring a Chinese word first.

Example concept: a due date.

Build direct routes to:

deadline
due date
by Friday
needs to be submitted
we’re running out of time

你看到“截止时间这个现实概念”，就能直接进入多个英语表达，而不是必须先出现“截止日期”四个中文字。

12. Learn meaning families, not translation pairs | 学“意义家族”，不要只学一对一翻译

Concept: uncertainty.

maybe
perhaps
probably
it seems
it appears
I’m not sure
I wouldn’t rule out

Each has different strength and register. Lesson 002 owns the deep calibration; here the goal is fast access to multiple routes.

13. Retrieval speed matters | 高级口语需要“提取速度”

A word you can remember after thirty seconds may still be unavailable for conversation.

Speaking vocabulary must be retrievable inside the time window of interaction.

训练方法：

topic cue → phrase within 2–3 seconds
function cue → discourse phrase
Chinese concept → multiple English options
English paraphrase when exact word fails

14. Paraphrase is an emergency route and a mastery route | 改述既是“救命技能”，也是高级能力

If you cannot retrieve unintended consequence, say:

a result that nobody planned

Then continue.

Do not stop the conversation while searching for the perfect phrase.

高级学习者会优先保护意义，再在之后升级语言。

15. The “simple first, precise second” strategy | “先简单说出来，再变精准”

Version 1:

It creates another problem.

Version 2:

More specifically, it shifts the cost onto smaller firms.

This keeps speech moving while allowing refinement.

你不需要第一毫秒就输出最终完美版本。

16. Reformulation is a normal advanced behaviour | 改口不是失败

CEFR C2 explicitly recognises smooth backtracking and reformulation as part of high-level spoken fluency.

高级说话人也会改口。差别是他们修复得很自然。

Useful repair phrases:

what I mean is…
let me put that differently…
actually, a better way to say it is…
more precisely…
that’s not quite what I mean…
let me correct that…

17. Self-repair is learning in public | 自我修正是“公开学习”

Research in Chinese EFL settings has examined self-repair as a way learners manage lexical gaps, errors and interactional problems. Repair is not merely evidence of weakness; it can be an active problem-solving process.

研究中国 EFL 课堂口语的工作指出，self-repair 可以被理解为学习者解决词汇空缺、语言错误和互动问题的一种行动。高级训练不应该要求“绝对不改口”，而应该训练“如何改得快、改得清楚、改完继续”。

18. Distinguish productive repair from destructive repair | 区分“有效修复”和“破坏流利度的修复”

Productive:

It was launched in May—sorry, June—and the first results came in August.

Destructive:

It was… it is… it has… no… it… launched… sorry…

The second speaker may be monitoring too many grammar possibilities at once.

高级口语要降低“过度实时编辑”。

19. Do not edit every error while speaking | 说话时不要修每一个错误

Repair errors that:

change meaning
create ambiguity
damage the main relationship
are easy to repair quickly

Let some small errors go if repairing them would destroy the larger message.

实时口语的目标不是现场变成 grammar checker。

20. Accuracy can be repaired after the turn | 有些准确度问题可以在说完后修

Record first. Review later.

This separates:

performance mode
learning mode

如果你同时要求“流利 + 复杂 + 零错误 + 完美发音”，工作记忆很容易过载。

21. Pauses are not all bad | 停顿不全是坏事

A pause at a clause boundary can help the listener. A long pause in the middle of a basic phrase often signals retrieval trouble.

Track pause location, not only pause count.

例如：

There are two main reasons. / First,…

这样的停顿有结构价值。

22. Replace silent collapse with discourse holding | 不要“沉默崩掉”，学会保持话轮

When you need time:

let me think for a second
the first thing that comes to mind is…
there are a couple of ways to look at that
I’d probably start with…

这些表达让你继续拥有话轮，同时进行规划。

23. But avoid filler addiction | 但不要依赖填充词

If every sentence begins:

Well, actually, basically, you know…

the filler stops helping.

Use discourse markers when they perform a job.

24. Think in “speech paragraphs” | 用“口语段落”思考

A strong one-minute answer often has:

position
reason
example
qualification
return

Example:

I think smaller classes can help, mainly because teachers can give more targeted feedback. For example, a writing teacher can notice recurring errors much faster when there are fewer scripts to review. That said, class size alone won’t solve weak teaching. So I’d treat it as an enabling condition, not a complete solution.

这就是口语里的“段落架构”。

25. One-minute answers should move | 一分钟回答要“有运动”

A weak answer circles:

I think it’s good because it’s good for people and there are many benefits.

A stronger answer develops:

position → mechanism → example → limit.

26. Compare, do not describe twice | 比较时不要分别描述两遍

Cambridge C1 long-turn tasks require comparison and speculation.

Weak:

In the first picture… In the second picture…

Stronger:

Both situations involve teamwork, but the first looks highly structured whereas the second seems more spontaneous.

比较是一种关系语言。

27. Speculation needs calibrated language | 推测要控制确定程度

They might be…
It looks as though…
They seem to be…
It’s possible that…
I’d guess that…

不要把“我不确定”变成“我什么都不敢说”。高级口语允许合理推测。

28. Justification turns opinion into reasoning | 给理由，观点才开始变成论证

Opinion:

I prefer hybrid work.

Justified:

I prefer hybrid work because it preserves some face-to-face collaboration without requiring daily commuting.

Advanced interaction repeatedly asks: why?

29. Use examples strategically | 例子要服务观点

Do not tell a long story that makes the listener forget your point.

Structure:

claim → short example → return to claim.

30. Interaction means building on what someone said | 互动不是轮流背答案

Weak:

Speaker A gives an answer. Speaker B gives an unrelated answer.

Strong:

I agree with your point about cost, and I’d add that implementation time matters as well.

高级 interaction 会把自己的话接到别人的话上。

31. Refer back explicitly | 明确连接前一个说话人

building on what you said…
I agree with the first part, but…
that’s a good point about X; I’m less sure about Y
the example you gave suggests…

这些表达帮助建立 joint discourse。

32. Turn-taking is a skill | 轮流说话是技能

Advanced speakers notice:

falling intonation
completed syntax
pause length
eye contact / non-verbal cues
discourse completion markers

Lesson 005 trained you to hear these cues. Lesson 006 trains you to act on them.

33. Enter a conversation without interrupting badly | 怎样自然进入对话

Useful entries:

Can I add something to that?
Just to build on that point…
I think there’s another issue here…
That’s true, although…

不要等到“完美空档”才永远不说，也不要直接压过别人。

34. Keep the floor when needed | 需要时要能保持话轮

There are two parts to this.
Let me finish that point.
Just one more thing.
The reason I mention this is…

CEFR C1 interaction descriptors explicitly include selecting phrases to get or keep the floor.

35. Yield the floor clearly | 说完以后让别人容易进来

Use:

What do you think?
Would you agree?
That’s how I see it, anyway.
So I’d probably choose option B.

互动能力包括“结束得让别人接得住”。

36. Backchanneling shows active listening | Backchannel 表示你在跟

right
exactly
mm-hm
I see
sure

但注意：right 有时表示“我听到了”，不一定表示“我同意”。

37. Clarification is advanced, not embarrassing | 请求澄清是高级沟通

Sorry, what do you mean by “scalable” here?
Do you mean cost in the short term or overall?
When you say “better”, better in what sense?
Could you give an example?

真正高级的互动，不是假装听懂。

38. Confirm meaning | 主动确认共同理解

So if I understand you correctly…
Just to check, you’re saying…
So the main concern is cost, not feasibility?

这会减少误解，也能给你时间整理下一步。

39. Disagreement has degrees | 反对有强度等级

I’m not sure I agree.
I see the logic, but I think the evidence points another way.
I’d question that assumption.
I don’t think that follows.
I strongly disagree with that conclusion.

Lesson 002 owns register calibration. Here the goal is real-time access to the right level.

40. Attack the idea, not the person | 反对观点，不攻击人

Instead of:

You’re wrong.

Try:

I’m not convinced that the evidence supports that conclusion.

这对学术讨论、职场会议和高阶口试都非常重要。

41. Agree partially | 学会“部分同意”

I agree up to a point.
I agree with the diagnosis, but not the solution.
That’s true in the short term, although…
I think that works for some cases, but…

高级交流很少只有全同意 / 全反对。

42. Negotiation is not persuasion alone | 协商不是只让别人接受你

Cambridge C1 collaborative tasks require reaching a decision together.

Negotiation includes:

propose
invite response
compare
concede
modify
trade off
decide

真正的目标是共同得到可行结果。

43. Proposal language | 提议语言

What if we…?
We could…
One option would be…
How about…?
It might make sense to…

44. Evaluation language | 评价选项

The advantage is…
The main drawback is…
That seems more practical because…
I’m not sure it solves the main problem.
It depends on the cost.

45. Decision language | 做决定

So shall we go with…?
On balance, I’d choose…
It sounds as though we agree on…
Let’s rule out X and compare Y with Z.

高级口语要能把“讨论”推进到“状态改变”。

46. Speak in contrasts | 用对比让思考更清楚

Useful frames:

not X but Y
the difference is that…
whereas…
in the short term…, but in the long term…
for individuals…, but at system level…

对比结构能快速生成高质量口语推理。

47. Speak in criteria | 用标准比较，不用 vague adjectives

Instead of:

Option A is better.

Say:

Option A is cheaper and faster, but Option B is more reliable.

这让讨论变得可协商。

48. Speak in mechanisms | 说“为什么会发生”

Instead of:

Small groups are better.

Say:

Small groups can help because the teacher can see more of each learner’s errors and respond faster.

这和 Lesson 004 的 reasoning system 直接连接。

49. Speak in conditions | 用条件让观点变精准

provided that…
as long as…
unless…
if the goal is…
assuming that…

高级观点不是“永远都对”，而是知道在什么条件下成立。

50. Think aloud safely | 学会“安全地边想边说”

In discussion:

My first reaction is…
I’d need to think about the cost, but…
One possibility is…
I’m not completely sure, but I suspect…

你不需要每次都等到最终结论才开口。

51. Mark uncertainty honestly | 不确定就准确表达不确定

Do not fake certainty to sound advanced.

Advanced speaking can say:

I don’t know enough about the technical side to be certain, but from a user perspective…

这叫 scope control。

52. Build a stance ladder | 建立立场强度阶梯

Strength	Possible language
tentative	I wonder if… / It might be…
moderate	I think… / It seems to me…
strong	I’m convinced that…
very strong	I strongly believe… / The evidence clearly shows… — only when justified

右边不是“更高级”，只是更强。

53. Use discourse markers to guide the listener | 让听者知道你现在在做什么

firstly
the main reason is
for example
that said
more importantly
to put it differently
so overall

口语 coherence 很大一部分来自这些路标。

54. Avoid connector performance | 不要为了“高级感”堆连接词

Do not force:

Moreover, furthermore, nevertheless, consequently…

unless the logic is real.

先有关系，再有连接词。

55. Do not speak like an essay | 不要把每次聊天都说成论文

Spoken English uses:

shorter clauses
repairs
discourse markers
contractions
ellipses
interactional phrases

高级不等于所有句子都复杂正式。

56. Spoken grammar can be simpler but discourse can be sophisticated | 口语语法可以简单，但互动可以高级

I’m not sure that works. The problem is cost. If the budget is fixed, we’d probably have to cut something else.

Grammar is not spectacular. Reasoning is clear.

高级口语的价值经常在 discourse control，而不是“每句都从句套从句”。

57. Complex sentences still matter—but should grow naturally | 复杂句要自然长出来

Start:

I support it.

Grow:

I support it because it reduces waiting time.

Grow:

I support it because it reduces waiting time, although I’d want to know how much it costs.

这比先在脑中设计一个 45 词完美复杂句更适合实时口语。

58. Pronunciation serves meaning | 发音服务意义

Do not pause speaking until every phoneme is perfect.

Priorities:

intelligibility
word stress
sentence prominence
thought grouping
key consonant/vowel contrasts that affect meaning

Lesson 005 trained perception; speaking now applies those patterns.

59. Stress key information | 重读真正重要的信息

The main problem is the COST.

I said TUESDAY, not Thursday.

Prominence helps listeners follow meaning.

60. Weak forms can improve natural flow | 弱读可以帮助自然语流

Function words do not need equal stress. But do not deliberately swallow sounds so aggressively that intelligibility drops.

自然度来自结构，不来自“故意说含糊”。

61. Thought groups make long answers easier | 意群让长回答更容易听

Speak:

There are two reasons / I’d support the proposal. / First, it reduces cost. / Second, it can be introduced gradually.

意义分组比不停不停地一口气说完更高级。

62. Pace should change with information density | 语速应该随信息密度变化

Slow slightly for:

numbers
technical terms
key distinctions
final decisions

真实高级沟通会主动保护重要信息。

63. Repetition can be useful | 重复有时是功能，不是错误

Repeat to:

emphasise
clarify
repair
hold the floor

But unconscious repeated fillers signal planning trouble.

64. Record yourself because memory is unreliable | 录音，因为你对自己口语的记忆不可靠

After speaking, you may remember the idea, not the exact language.

Recording reveals:

pause location
repeated words
repair patterns
grammar under pressure
pronunciation problems
idea structure

65. Use a speaking transcript selectively | 选择性转录自己的口语

You do not need to transcribe 20 minutes.

Choose 60–90 seconds and mark:

pauses
fillers
restarts
vague vocabulary
good phrases worth keeping

一小段就能暴露很多系统问题。

66. Measure pause location | 记录“在哪里停”

If you often pause:

before nouns → lexical retrieval
before every sentence → planning structure
inside fixed phrases → chunk automation weak
after grammar starts → over-monitoring

停顿位置比“我总是卡”更可诊断。

67. Measure repetition type | 分析重复的类型

I think, I think, I think… may show launch failure.

The main issue—the main issue—is cost. may be emphasis or repair.

不要把所有重复归为同一问题。

68. Build a speaking error corpus | 建立个人口语错误库

Symptom	Cause	Repair
long pause before every answer	full-sentence translation	English node planning
same simple words	active retrieval weak	topic phrase sets
many restarts	over-monitoring	simple-first strategy
grammar collapses in long turns	complexity overload	incremental sentence growth
speech sounds like essay	written register transfer	conversation model analysis

69. Repeated speaking is one of the highest-value drills | 同一个任务重复说，是高价值训练

Use:

speak once
review
repair 3–5 features
speak again
wait 24–48 hours
speak again without notes

第一次测现状，第二次做修复，第三次看修复是否进入自动化。

70. Do not change topic too quickly | 不要每次练习都换新话题

New topics constantly force conceptual planning, making language improvement harder to see.

Use the same topic across several sessions and change the speaking job:

explain
argue
compare
persuade
summarise

71. The 4/3/2 technique can train fluency | 4/3/2 可以训练流利度

Explain the same content in:

4 minutes
3 minutes
2 minutes

The content remains similar while time pressure increases efficiency.

不要把它变成纯“说更快”。目标是减少不必要停顿、提高词块提取和信息压缩。

72. Retelling links listening to speaking | 复述把 Lesson 005 接到 Lesson 006

Listen to a 3–5 minute segment, then retell:

30 seconds
2 minutes
full version

同一材料训练理解、检索和实时组织。

73. Opinion drills should include reasons and limits | 观点训练必须带理由和限制

Prompt:

Should universities record all lectures?

Answer architecture:

position → two reasons → one concern → final judgement.

74. Unknown-topic practice builds real-time thinking | 陌生话题训练实时思考

You do not need specialist knowledge to discuss every issue intelligently.

Use:

define what you know
state uncertainty
reason from general principles
ask clarifying questions

这连接 eduKate 的 unfamiliar-topic oral owners，但这里应用到 C1–C2 成人/高级场景。

75. Learn to say “I don’t know” at C1 level | 高级英语也要会准确说“不知道”

Weak:

I don’t know.

More useful:

I don’t know enough about the technical side to answer confidently, but I can see two practical issues.

这不是逃避，是 scope management。

76. Use time strategically | 实时回答也需要“时间管理”

For a one-minute response:

10 sec: position
20 sec: reason 1
20 sec: reason 2/example
10 sec: qualification/conclusion

不要在前 45 秒只铺背景。

77. Speaking anxiety changes performance | 焦虑会改变真实口语表现

Chinese EFL research has examined the roles of self-efficacy, enjoyment, anxiety and willingness to communicate in oral performance. A learner can possess language knowledge but underperform when interaction is associated with excessive monitoring or threat.

研究中国 EFL 学习者的工作显示，self-efficacy、enjoyment、anxiety、willingness to communicate 都和口语发展/表现有关。语言知识存在，不代表压力下能稳定调用。

78. Reduce anxiety by reducing task uncertainty | 降低任务不确定性，而不是等“完全不紧张”

Before practice, define the job:

explain one idea
compare two options
defend one position
ask three follow-up questions

清楚任务边界，会减少“我不知道要说什么”的焦虑。

79. Safe repetition builds automaticity | 安全重复建立自动化

Repeat familiar speaking tasks privately before increasing social difficulty.

Progression:

self-recording
AI / tool practice if useful
trusted partner
small group
unfamiliar interlocutor
high-stakes presentation

这不是逃避真实环境，而是渐进负荷。

80. Speaking assistants can increase practice volume—but do not outsource judgement | 口语工具可以增加练习量，但不要外包判断

Recent Chinese EFL studies have reported benefits from structured technology-supported speaking practice for willingness to communicate, enjoyment or oral performance. But tool feedback can be imperfect.

Use technology for:

practice frequency
recording
transcription
prompt generation
comparison across attempts

Do not assume every automated correction is linguistically or pragmatically right.

81. Grammar knowledge must survive forward planning | 高级语法要在“一边想下一句”时仍然稳定

CEFR C2 accuracy explicitly recognises maintaining grammatical control even while attention is engaged in forward planning or monitoring others’ reactions.

这解释了高级语法为什么不能只做 worksheet：真正测试发生在你同时思考内容、听对方、规划下一步时。

82. Simplify grammar under pressure | 压力大时先简化结构

Instead of planning:

Had the policy been implemented earlier, it might conceivably have prevented…

if unstable, say:

If the policy had started earlier, it might have prevented…

稳定表达优先于炫技。

83. Complexity should be added after control | 先控制，再加复杂度

Use three stages:

clear simple answer
add one qualification
add one contrast/condition

高级复杂度应该是“可控增加”。

84. Translate functions, not sentences | 实时互动要翻“功能”，不要翻整句

Chinese function: politely disagree.

Do not build the Chinese sentence first. Activate English function routes:

I’m not completely convinced.
I see the point, but…
I’d question whether…
I think the evidence points another way.

这就是从“句子翻译”升级到“交际功能调用”。

85. Build function banks | 建立功能词库

Create banks for:

agree
disagree
clarify
repair
compare
speculate
recommend
negotiate
summarise
close a discussion

每个功能存 5–10 条不同强度、不同语域表达。

86. Build topic banks | 建立话题词块，而不是整篇稿

Topic: education.

Keep:

learning outcome
feedback quality
independent performance
teacher workload
access gap

不要背完整答案。背“可重组部件”。

87. Build connector banks | 建立逻辑连接词库

because
so
but
although
whereas
that said
in other words
as a result
provided that

真实口语里，简单连接词非常有价值。

88. Build repair banks | 建立修复词库

sorry, let me correct that
what I mean is
more precisely
I should have said
let me rephrase that

89. Build interaction banks | 建立互动词库

what do you think?
would you agree?
can I add something?
that’s a fair point
can you clarify what you mean by…?

90. Learn the sound of your chunks | 词块必须进入声音系统

A phrase stored only visually is slow in speech.

For each useful phrase:

listen
mark stress
say it
use it in a new sentence
retrieve it later

Lesson 005 and Lesson 006 should share the same phrase inventory.

91. Use shadowing for launch phrases | 用 shadowing 训练起步词块

Shadow short discourse phrases until they can be produced without visual support.

This reduces launch latency.

92. Use substitution drills intelligently | 替换练习也可以高级

Frame:

The main reason is that ___.

Replace content rapidly:

cost
time
risk
learning quality

目标不是机械语法，而是自动化 discourse frame。

93. Use transformation drills | 用变换训练灵活性

Say the same idea:

casually
professionally
academically
in 20 seconds
in 60 seconds

这同时训练 register、compression 和 reformulation。

94. Use forced paraphrase | 强制改述训练

Ban one key word and explain the idea anyway.

Example: explain inflation without saying inflation.

This makes semantic access stronger than lexical dependence.

95. Use interruption drills carefully | 模拟被打断后的恢复

Have a partner interrupt with:

why?
can you give an example?
what do you mean?
I disagree.

Then return to your main line.

高级口语不是只有 prepared monologue。

96. Use surprise questions | 加入意外追问

After a prepared answer, ask an unprepared follow-up.

This tests whether the learner owns the idea or only memorised the wording.

97. Use decision tasks | 用共同决策训练互动

Choose three options and require the pair to agree on one.

Track:

idea exchange
response to partner
concession
criteria
decision language

98. Use role switching | 换立场训练灵活性

Argue for a position for two minutes, then argue against it.

This prevents vocabulary from becoming tied to one memorised opinion.

99. Use audience switching | 换对象训练语域

Explain the same idea to:

a friend
a manager
a teenager
an academic audience

这连接 Lesson 002。

100. Use one-topic, many-functions practice | 一个话题，多种交际功能

Topic: public transport.

describe
compare
complain
recommend
negotiate
summarise

真正灵活的口语不是“很多话题，每个只会一句”。

101. A 20-minute daily speaking routine | 每天 20 分钟实时口语

Time	Task
5 min	retrieve 8–10 function/chunk phrases
5 min	one 2-minute no-note answer + review
5 min	repeat answer with 3 repairs
5 min	one surprise follow-up

102. A 45-minute growth session | 45 分钟增长模式

Time	Task	Focus
8 min	chunk/function retrieval	access speed
10 min	repeated long-turn task	fluency
10 min	interaction/decision task	turn-taking
10 min	recording review + repair	feedback
7 min	delayed no-note rerun	retention

103. A 90-minute deep session | 90 分钟深度口语

Time	Task
15 min	input + phrase mining from authentic dialogue
15 min	shadowing + prosody
20 min	long-turn speaking
20 min	collaborative discussion / negotiation
10 min	transcript/error analysis
10 min	final repeat

104. Seven-day real-time speaking cycle | 七天实时口语循环

Day 1: English node planning + launch phrases.
Day 2: chunks + retrieval speed.
Day 3: repeated speaking + repair.
Day 4: interaction + turn-taking.
Day 5: disagreement + negotiation.
Day 6: unfamiliar topic + surprise questions.
Day 7: benchmark recording + error corpus.

105. Twelve-week C1–C2 speaking progression | 12 周 C1–C2 口语路线

Weeks	Focus	Evidence
1–2	baseline + translation dependency diagnosis	recordings + pause map
3–4	chunks + direct concept routes	retrieval sets
5–6	long turns + reformulation + self-repair	repeated recordings
7–8	turn-taking + disagreement + clarification	pair discussions
9–10	negotiation + unfamiliar topics + register	decision tasks
11–12	independent C1–C2 discussion + transfer	final speaking portfolio

106. Monthly benchmark | 每月基准测试

Complete four tasks:

2-minute personal opinion
1-minute comparison/speculation
5-minute collaborative discussion
3-minute unfamiliar-topic response

Measure:

launch latency
pause location
lexical variety
grammar stability
repair quality
interaction
coherence
pronunciation intelligibility

107. The first weak link table | 实时口语第一个薄弱环节

Symptom	Likely first weak link	Repair
I know what to say but start slowly.	full Chinese sentence planning	node planning + launch phrases
I repeat basic words.	active lexical access	topic/function chunks
I restart sentences repeatedly.	over-monitoring	simple-first + delayed correction
I speak well alone but poorly with people.	interaction load	pair negotiation + turn-taking
I freeze on unknown topics.	idea-generation strategy	criteria/mechanism/condition reasoning
I sound too formal.	written-register transfer	spoken model comparison
People ask me to repeat.	intelligibility/prosody	stress + thought groups + key sounds

108. What not to do | 不要这样练高级口语

Do not memorise full essays as speeches. 不要背整篇作文当口语。
Do not ban Chinese thinking completely. 不要把母语变成敌人。
Do not wait for a perfect sentence before speaking. 不要等整句完美才开口。
Do not correct every small error in real time. 不要让自我监控摧毁主线。
Do not chase speed alone. 快不等于流利。
Do not practise only monologues. 互动才是真实口语。
Do not use only familiar topics. 需要迁移。
Do not replace interaction with scripts. 协商不是背答案。
Do not chase a native accent as the main target. 可懂度和意义更重要。

109. Research floor | 研究基础

Cambridge C1 Advanced Speaking Format — long turn, collaborative discussion, negotiation, opinion, speculation and interaction.
Council of Europe CEFR Spoken Language Qualitative Aspects — range, accuracy, fluency, interaction and coherence at C1–C2.
National study of Chinese EFL learners — speaking challenges including L1 interference and limited practice.
System (2024): speaking CALF development in Chinese EFL learners — longitudinal oral-task performance.
System (2024): lexical bundles and L2 oral fluency — multiword sequences and speed/breakdown/repair fluency.
Self-repair in Chinese EFL learners’ speech — self-repair as problem solving in spoken L2 use.
Heliyon (2024): oral proficiency and motivational regulation in Chinese EFL learners.

Research identifies useful mechanisms and recurring challenges. Your own recordings remain the final diagnostic evidence.

110. Canonical eduKate routes | eduKate 内部主页面

111. SEO language map | 本课自然覆盖的搜索意图

This lesson naturally serves learners searching for advanced English speaking, C1 English speaking, C2 English speaking, speak English without translating, think in English, English speaking for Chinese speakers, English speaking for Mandarin speakers, improve English fluency, English speaking fluency, spontaneous English speaking, speak English naturally, English self-repair, English turn-taking, English discussion skills, English negotiation language, advanced oral English, 高级英语口语, C1 英语口语, C2 英语口语, 英语流利度, 不翻译说英语, 怎样用英语思考, 中文母语英语口语, 英语口语怎么提高, 英语实时表达, 英语讨论, 英语协商 and 英语改述.

关键词只是入口。本课真正拥有的是：高级学习者如何在实时互动中让“意义 → 英语 → 互动 → 修复”持续运行。

112. The one-page real-time speaking operating system | 一页实时口语操作系统

Think in concepts, not full Chinese sentences. 用概念，不先造完整中文句。
Plan with English nodes. 用 position / reason / example / limit。
Launch early. 用安全起步词块先开口。
Build incrementally. 句子边说边长。
Use chunks. 高频交际功能预编译。
Paraphrase when blocked. 找不到完美词也要继续。
Repair meaning-changing errors. 修重要错误，不修所有小错。
Interact. 回应、澄清、协商、接别人观点。
Record and diagnose. 用真实录音建立错误库。
Repeat after repair. 同任务再次输出，让修复进入自动化。

113. Final assignment | 本课最终作业

Choose one serious but familiar topic. Do not prepare a script.

选择一个你熟悉但有思考价值的话题。不要写完整稿。

Complete this sequence:

Write only five English nodes: position, reason 1, reason 2, example, limitation.
Speak for two minutes without Chinese notes.
Listen to the recording.
Mark every pause longer than two seconds.
Classify each pause: idea, vocabulary, grammar, pronunciation or monitoring.
Choose five better chunks or reformulation routes.
Speak again immediately.
Ask another person or tool to challenge one part of your view.
Respond without preparing.
Wait 48 hours.
Speak on the same topic again with no notes.

Then ask:

Did I speak faster because I forced myself to hurry?

That is not the target.

Did I speak more continuously because the route from meaning to English became shorter?

That is the target.

真正的进步，不是嘴巴更急，而是“意义 → 英语”的路线更短，互动和修复越来越自动。

Continue | 继续

The next lesson isolates one part of spoken English that deserves its own system: pronunciation for intelligibility, word stress, rhythm, thought groups and intonation.

下一课把口语里一个值得独立训练的系统拆出来：为了可懂度而训练发音、单词重音、节奏、意群和语调。

Lesson No.007 · Speak Clearly With Intelligibility, Stress, Rhythm and Intonation · 用可懂度、重音、节奏与语调说清英语

Back to Advanced English Chinese Edition Hub · 返回高级英语中文版主页

Small Group Tutorials

How to Learn Advanced English (Chinese Edition) | Lesson No.006 | Speak Without Translating in Real Time | 第006课：实时英语表达：不用先翻译再开口

Speak Without Translating in Real Time | 实时英语表达：不用先翻译再开口

Before we begin: this lesson owns real-time production | 开始之前：本课负责“实时口语生成层”

1. C1–C2 speaking is interaction, not a monologue contest | C1–C2 口语不是“连续说很久比赛”

2. Fluency has several parts | 流利度不是只有“语速”

3. Chinese learners themselves report the translation bottleneck | 中国学习者自己也报告“先中文再翻译”瓶颈

4. Do not ban Chinese thinking | 不要禁止中文思考

5. Separate concept from sentence | 把“概念”与“句子”分开

6. Build English from nodes | 用英语节点建句

7. Speak before the sentence is complete | 不要等整句完美以后才开口

8. Launch phrases reduce planning pressure | 起步词块可以降低实时规划压力

9. Lexical bundles can support oral fluency | 词块可以支持口语流利度

10. But do not turn chunks into scripts | 词块不是固定脚本

11. Build direct concept-to-English routes | 建立“概念 → 英语”直接路线

12. Learn meaning families, not translation pairs | 学“意义家族”，不要只学一对一翻译

13. Retrieval speed matters | 高级口语需要“提取速度”

14. Paraphrase is an emergency route and a mastery route | 改述既是“救命技能”，也是高级能力

15. The “simple first, precise second” strategy | “先简单说出来，再变精准”

16. Reformulation is a normal advanced behaviour | 改口不是失败

17. Self-repair is learning in public | 自我修正是“公开学习”

18. Distinguish productive repair from destructive repair | 区分“有效修复”和“破坏流利度的修复”

19. Do not edit every error while speaking | 说话时不要修每一个错误

20. Accuracy can be repaired after the turn | 有些准确度问题可以在说完后修

21. Pauses are not all bad | 停顿不全是坏事

22. Replace silent collapse with discourse holding | 不要“沉默崩掉”，学会保持话轮

23. But avoid filler addiction | 但不要依赖填充词

24. Think in “speech paragraphs” | 用“口语段落”思考

25. One-minute answers should move | 一分钟回答要“有运动”

26. Compare, do not describe twice | 比较时不要分别描述两遍

27. Speculation needs calibrated language | 推测要控制确定程度

28. Justification turns opinion into reasoning | 给理由，观点才开始变成论证

29. Use examples strategically | 例子要服务观点

30. Interaction means building on what someone said | 互动不是轮流背答案

31. Refer back explicitly | 明确连接前一个说话人

32. Turn-taking is a skill | 轮流说话是技能

33. Enter a conversation without interrupting badly | 怎样自然进入对话

34. Keep the floor when needed | 需要时要能保持话轮

35. Yield the floor clearly | 说完以后让别人容易进来

36. Backchanneling shows active listening | Backchannel 表示你在跟

37. Clarification is advanced, not embarrassing | 请求澄清是高级沟通

38. Confirm meaning | 主动确认共同理解

39. Disagreement has degrees | 反对有强度等级

40. Attack the idea, not the person | 反对观点，不攻击人

41. Agree partially | 学会“部分同意”

42. Negotiation is not persuasion alone | 协商不是只让别人接受你

43. Proposal language | 提议语言

44. Evaluation language | 评价选项

45. Decision language | 做决定

46. Speak in contrasts | 用对比让思考更清楚

47. Speak in criteria | 用标准比较，不用 vague adjectives

48. Speak in mechanisms | 说“为什么会发生”

49. Speak in conditions | 用条件让观点变精准

50. Think aloud safely | 学会“安全地边想边说”

51. Mark uncertainty honestly | 不确定就准确表达不确定

52. Build a stance ladder | 建立立场强度阶梯

53. Use discourse markers to guide the listener | 让听者知道你现在在做什么

54. Avoid connector performance | 不要为了“高级感”堆连接词

55. Do not speak like an essay | 不要把每次聊天都说成论文

56. Spoken grammar can be simpler but discourse can be sophisticated | 口语语法可以简单，但互动可以高级

57. Complex sentences still matter—but should grow naturally | 复杂句要自然长出来

58. Pronunciation serves meaning | 发音服务意义

59. Stress key information | 重读真正重要的信息

60. Weak forms can improve natural flow | 弱读可以帮助自然语流

61. Thought groups make long answers easier | 意群让长回答更容易听

62. Pace should change with information density | 语速应该随信息密度变化

63. Repetition can be useful | 重复有时是功能，不是错误

64. Record yourself because memory is unreliable | 录音，因为你对自己口语的记忆不可靠

65. Use a speaking transcript selectively | 选择性转录自己的口语

66. Measure pause location | 记录“在哪里停”

67. Measure repetition type | 分析重复的类型

68. Build a speaking error corpus | 建立个人口语错误库

69. Repeated speaking is one of the highest-value drills | 同一个任务重复说，是高价值训练

70. Do not change topic too quickly | 不要每次练习都换新话题

71. The 4/3/2 technique can train fluency | 4/3/2 可以训练流利度

72. Retelling links listening to speaking | 复述把 Lesson 005 接到 Lesson 006

73. Opinion drills should include reasons and limits | 观点训练必须带理由和限制

74. Unknown-topic practice builds real-time thinking | 陌生话题训练实时思考

75. Learn to say “I don’t know” at C1 level | 高级英语也要会准确说“不知道”

76. Use time strategically | 实时回答也需要“时间管理”