本课目标

学完本课，你能够：

• 在证据不完整时区分可能、较可能和不能排除；
• 提出替代解释而不假装每种解释同样有证据；
• 说明结果可能依赖对象、过程、时间或测量方式；
• 避免用maybe模糊带过全部判断；
• 把绝对结论改写成有概率、有条件、可继续核查的结论。

本课边界

继续使用第69课的虚构来源包。新增的八周记录也由教材原创，只用于语言训练。概率词表达的是对当前材料的把握，不提供真实统计概率；likely不等于“超过某个固定百分比”，除非材料明确给出模型和数值。

场景词汇

• alternative explanation 替代解释
• confounding change 同时发生、影响解释的变化
• baseline 基线
• week-to-week variation 周际波动
• measurement choice 测量选择
• plausible 合理可能的
• rule out 排除
• conditional conclusion 条件式结论

---

More than one story fits the same number

Maya and Marcus now have an eight-week action-item log and two short notes from the fictional support team. They test how many explanations the same trend could support.

Part A — Ranking explanations

Marcus: The weekly overdue counts are ten, nine, eight, nine, seven, seven, six and six. The evidence seems to suggest improvement, but not a smooth decline.

Maya: Agreed. It’s possible that shorter check-ins helped the team focus, especially because the team lead noticed that effect.

Marcus: One likely explanation is the new owner list. Every action had one named person after week two, so unfinished work became easier to follow up.

Maya: Why call that likely rather than simply possible?

Marcus: The notes connect two recovered items directly to owner clarification. That is still limited evidence, but it links a mechanism to specific cases.

Maya: We can’t rule out the written agenda either. The team began circulating it before meetings in week three, and the count fell again afterward.

Marcus: That doesn’t necessarily mean the agenda caused the fall. Week four rose from eight to nine even though the agenda was already in use.

Maya: True. The result may depend on how consistently the team used all three changes, not on whether a document technically existed.

Marcus: There is also a workload note: incoming requests were lower in weeks seven and eight.

Maya: Then lower workload is a plausible contributor to the final count of six. Do we know whether the size of the action items changed?

Part B — Making the conclusion conditional

Marcus: No. The log counts items, not effort, urgency or project importance. Six large overdue items could represent more work than ten small ones.

Maya: So the measurement choice matters. What can we still say with reasonable confidence?

Marcus: Over eight weeks, the recorded count fell from ten to six while meeting length, agenda use, ownership and follow-up changed. That trend is directly documented.

Maya: And what is more tentative?

Marcus: The shorter format may have contributed, but the evidence cannot separate its effect from the other changes or lower late-period workload.

Maya: Could the first week have been unusually bad?

Marcus: We can’t rule that out because there is no earlier baseline. Comparing only week one and week eight may exaggerate a normal fluctuation.

Maya: That may be true. However, the middle weeks mostly move downward, so the pattern is not based on the endpoints alone.

Marcus: A more accurate conclusion is: this team’s bundled process change was followed by fewer recorded overdue items, and shorter meetings are one plausible contributor.

Maya: Good. We can revisit causal weight only if we obtain a longer baseline, stable workload measures or a comparison that changes one feature at a time.

Eight-week fictional log

Week	Overdue items	Process note	Workload note
1	10	30-minute check-in; no standard agenda or owner list	not recorded
2	9	15-minute check-in; owner list begins	typical, according to team lead
3	8	written agenda added	typical
4	9	same process	two urgent requests
5	7	next-day review used consistently	typical
6	7	same process	typical
7	6	same process	lower incoming requests
8	6	same process	lower incoming requests

The log does not record item size, effort, urgency, earlier baseline weeks or a comparison team.

Check the story

1. Why is the downward pattern not smooth?
2. Why does Marcus view the owner list as a relatively likely explanation?
3. What happened in week four?
4. Which late-period condition could contribute to the final count?
5. What does the log measure and fail to measure?
6. What conditional conclusion do Maya and Marcus accept?

---

不确定不等于“什么都不知道”

当前材料能直接支持：

The recorded count fell from ten to six during a bundled process change.

只能有限支持：

Shorter meetings may have contributed.

不能支持：

Shorter meetings alone caused exactly forty percent improvement.

成熟的保留意见会把这三层分开。

替代解释需要排序

把十种解释全部列成maybe并无帮助。可按证据排序：

• 有具体机制和案例连接：较可能；
• 时间上同时出现但无直接连接：合理可能；
• 与材料矛盾或无任何迹象：当前较不可能，但未必完全排除。

指标决定故事

只数item会忽略大小、难度和重要性。一个结果看似改善，可能部分来自测量口径。核查时要问：What exactly was counted? What was not counted?

---

本课负荷说明

本课主动掌握6个新词块。CH399—CH404在练习中继续获得近R，L69来源语言也会再次调用。

1. The evidence seems to suggest …

给出有根据但有限的方向判断；分析性。

The evidence seems to suggest a downward trend.

2. It’s possible that …

提出合理可能，不暗示它最有支持。

It’s possible that shorter meetings helped maintain focus.

3. One likely explanation is …

指出相对更有材料支持的解释；必须说明依据。

One likely explanation is clearer ownership, because two cases were recovered after owners were named.

4. We can’t rule out …

说明当前材料不足以排除，不等于认可或认为概率很高。

We can’t rule out an unusually high first week.

5. That doesn’t necessarily mean …

阻止从共同发生跳到必然因果。

The count fell after the agenda appeared; that doesn’t necessarily mean the agenda caused the change.

6. The result may depend on …

指出效果可能随条件变化。

The result may depend on workload and consistent follow-up.

Don’t Say It That Way

容易说出的句子	问题	更准确的表达
Maybe anything caused it.	没有排序、机制或证据	Several explanations remain; ownership has the clearest case link.
It was probably the meetings.	probably无依据	Shorter meetings may have contributed, but the effects cannot be separated.
We can’t rule it out, so it happened.	把未排除写成确认	We cannot rule it out; current evidence does not show its weight.
The number went down, so performance improved.	指标过窄	Recorded overdue items fell; item size and overall performance were not measured.

---

1. 从较弱到较强的常用梯度

could be / might be / may be → is possible → is plausible → is likely

这不是固定数学刻度。使用likely时必须给出比“想得到”更强的理由。

2. 过去可能原因：may/might/could have + 过去分词

Shorter meetings may have helped.
Lower workload could have contributed.
Week one might have been unusually high.

它们表达对已发生事件的推测，不是情态完成式责备。

3. must have / can’t have要求更强证据

The owner list must have caused the decline.

当前材料无法排除其他解释，因此must have过强。can’t have同样需要强排除依据。

4. not necessarily拆开“发生”与“必然含义”

Counts fell after the agenda began. That doesn’t necessarily mean the agenda caused the fall.

不是说因果一定不存在，而是说当前观察不足以证明。

5. 条件限制概率

The format may help when ownership is clear.
The result may depend on consistent use and workload.

条件越具体，结论越能被检验。

快速练习

选择：

1. Lower workload (may have / must have) contributed; the log cannot separate its effect.
2. One likely explanation ______ clearer ownership. (is / are)
3. We can’t rule ______ an unusually high first week. (out / off)
4. A fall after week three does not necessarily ______ the agenda caused it. (mean / means)
5. The result may depend ______ item size and workload. (on / of)
6. The middle weeks (support / prove) a downward trend but not a single cause.

---

概率词通常轻读，重音落在解释内容：

One likely explanation is / CLEARER OWNERSHIP.

can’t rule out重读不能排除的因素，不读成确认：

We can’t rule out / a HIGH FIRST WEEK.

修正因果时，在necessarily与mean之间保持连贯：

That doesn’t necessarily mean / the agenda CAUSED IT.

影子跟读

The evidence seems to suggest a downward trend.
One likely explanation is clearer ownership.
We can’t rule out lower workload.
That doesn’t necessarily mean the meeting length had no effect.

---

A. 理解检查

判断正误：

1. The overdue count fell every week.
2. Two recovered items were linked to clearer ownership.
3. The agenda began before the shorter meetings.
4. Incoming requests were lower in weeks seven and eight.
5. The log records the size and urgency of every item.
6. Maya and Marcus rule out any role for shorter meetings.

B. Direct / Plausible / Unsupported

分类：

1. The recorded count was ten in week one and six in week eight.
2. Shorter meetings may have contributed to focus.
3. The owner list was the only cause.
4. Lower workload could explain part of the final decline.
5. Overall performance improved by exactly forty percent.
6. The count rose in week four.

C. 给解释排序并说明依据

对以下解释使用possible / plausible / likely / cannot assess，每项写一条依据：

• shorter meetings;
• named owners;
• written agenda;
• next-day review;
• lower workload;
• smaller action items;
• unusually high week one.

D. CH399—CH404第二次近R

回应Short meetings always make teams more productive，必须加入来源、指标和条件，不得只把always换成sometimes。

E. Lesson 69检索

用CH405—CH410追溯“40%改善”说法，再分别标记：帖子说法、通讯摘要、客户案例、尚未找到的证据和下一追溯问题。

F. 概率会议

三人分别主张短会议、owner list和低工作量是主要解释。完成12—15句：

• 每人给一个机制和一条材料；
• 承认一个替代解释；
• 指出一项不能排除的因素；
• 写出共同接受的条件式结论；
• 不进行多数投票决定“真相”。

G. 非课文阅读：两条补充记录

Note A: “In week four, two urgent customer requests created six extra action items. The meeting process did not change.”
Note B: “In weeks seven and eight, incoming requests were lower than usual. No effort estimate was recorded.”

回答：两条记录分别提高或降低了哪些解释的可信度？它们仍不能回答什么？

H. 3分钟条件式结论录音

用本课记录完成：

1. 直接趋势；
2. 最可能解释及理由；
3. 两个合理替代解释；
4. 一个不能排除的测量问题；
5. 一个依赖条件；
6. 一句最有限结论和下一数据需求。

I. 实用写作：解释矩阵

写160—210词，比较至少四种解释。每种写：机制、支持材料、反对材料、当前把握和所需新证据。结尾不得使用proved / definitely / no doubt。

J. 中译英情境表达

1. 证据似乎表明记录中的逾期项呈下降趋势。
2. 较短的会议有可能帮助维持专注。
3. 一个较可能的解释是责任人更清楚。
4. 我们不能排除第一周异常偏高。
5. 这不一定意味着书面议程造成了下降。
6. 结果可能取决于工作量和执行的一致性。

---

建立解释矩阵

解释	可能机制	支持	反证／限制	当前用词
				possible / plausible / likely / unknown

至少填写四行。若每行都写possible，说明你尚未比较证据强弱。

禁用maybe重录

第一遍允许自由说；第二遍禁止使用maybe，改用本课6个词块和具体条件。目标不是显得正式，而是让听者知道你有多大把握、为什么。

三天后的复测

只看八周表，3分钟给出两种解释及条件式结论。若把cannot rule out说成confirmed，重新复习证据强度。

---

• 我能区分直接记录、合理解释和无支持结论。
• 我能为likely提供比“想得到”更强的依据。
• 我能正确使用may/might/could have讨论过去可能原因。
• 我不会把cannot rule out误写成“已经证实”。
• 我能说明指标、基线和工作量如何影响解释。
• 我完成了3分钟条件式结论、160—210词解释矩阵和两组近R。
• 我不会给虚构材料赋予真实统计概率或工作建议地位。