# 英语中级 · 第 9 周 Day 4|三句摘要初稿 > 系列:语种学习 / 英语 / 中级 · 每日学习|AI生成|2026-10-27|codebuddy **今天学什么** 今天第一次把一周的输入压成 100 词。三句摘要法的好处不是省事,是逼你排序:哪一句是文章的主张,哪一句是支撑它的证据,哪一句是作者希望你带走的影响。三句写不出来,通常不是英语不够,是你还没想清文章的结构。今天的产出是初稿,允许粗糙,但结构必须对——三句各干一件事,一句都不许多。写之前先列三行关键词提纲,列不出来就回去看 Day 1 的主题句。 ## 一、今日任务 | 环节 | 时长 | 具体动作 | 产出 | | --- | --- | --- | --- | | 列三行提纲 | 8 min | 只写关键词,不写完整句:主旨一行、证据一行、影响一行;每行不超过 8 个词 | 3 行提纲 | | 写第 1 句 | 12 min | 用 The article argues that… 起句,30—40 词,写清「对谁、往哪个方向、主张什么」 | 主旨句 | | 写第 2 句 | 15 min | 用 It supports this by… 或 Drawing on…,点名证据类型与关键数字,不抄原句 | 支撑句 | | 写第 3 句 | 12 min | 用 The implication is that… 收尾,只写文章自身的结论,不添自己的看法 | 结论句 | | 压缩与朗读 | 13 min | 数词数压到 100 词内,先删形容词副词、再删重复、最后删例证细节;然后出声读一遍,卡住处拆句或换词 | 100 词内初稿 + 朗读标记 | 100 词不是「差不多就行」,按空格数,连字符算一个词,标题和括号里的词数标注不算。写到 115 词也别慌,先保证三句结构,Day 5 再压缩——初稿的任务是把话说全,不是一次写完美。 60 分钟里最该守住的是「先提纲后句子」。直接写完整句的人,十有八九写到第三句发现第一句已经把所有内容说完了,于是第三句变成重复。提纲只有三行,改起来十秒钟;三句写完再改结构,要花二十分钟。 ## 二、核心词汇与搭配 | 词 | 音标 | 词性 | 释义 | 高频搭配 | 例句(含翻译) | | --- | --- | --- | --- | --- | --- | | argue | /ˈɑːɡjuː/(≈阿格尤) | v. | 主张、论证 | argue that …;argue for | The article **argues** that the fee should be capped.(文章主张收费应设上限。) | | maintain | /meɪnˈteɪn/(≈梅恩泰恩) | v. | 坚持认为 | maintain that … | The author **maintains** that the trend is temporary.(作者坚持认为这一趋势是暂时的。) | | support | /səˈpɔːt/(≈瑟保特) | v. | 支持、支撑(论点) | support this by;be supported by | The claim is **supported by** two case studies.(这一主张由两个案例支撑。) | | draw on | /drɔː ɒn/ | phr. v. | 借助、取材于 | draw on data / research | The author **draws on** ten years of tax records.(作者取材于十年税务记录。) | | demonstrate | /ˈdemənstreɪt/(≈戴默恩斯特雷特) | v. | 证明、演示 | demonstrate that … | The experiment **demonstrates** that the rule works.(实验证明该规则有效。) | | conclude | /kənˈkluːd/(≈克恩克卢德) | v. | 得出结论 | conclude that … | The report **concludes** that costs will fall.(报告得出结论:成本将下降。) | | implication | /ˌɪmplɪˈkeɪʃn/(≈伊姆普利凯希恩) | n. | 含义、影响 | the implication is that;implications for | The **implication** is that prices must rise.(这意味着价格必须上涨。) | | overall | /ˌəʊvərˈɔːl/(≈欧弗劳尔) | adv. / adj. | 总体而言 | overall, …;overall trend | **Overall**, the evidence points one way.(总体而言,证据指向同一方向。) | | point to | /pɔɪnt tu/ | phr. v. | 指向、表明 | point to a decline | The figures **point to** a deeper problem.(数字指向一个更深的问题。) | | in short | /ɪn ʃɔːt/ | phr. | 简言之 | In short, … | **In short**, the policy helps few and costs much.(简言之,这政策帮的人少、代价大。) | 四个转述动词的力度不一样:argue 是作者主动站队并给理由,maintain 是面对质疑仍坚持,claim 暗含「未必可信」,show 是证据已经摆出来了。摘要第一句用 argue 最稳,因为评论文本来就在论证;如果你读的是报道而非评论,改用 reports 或 finds,不要硬套 argue。 写摘要别用 "The article talks about"、"It is very important"。这两个说法在 B1 阶段就该淘汰:前者没有立场,后者没有信息。换成 argue、point to、show,句子立刻有骨头。 ## 三、语篇点:三句各干一件事 **五条规则** 1. 第 1 句只负责主旨:谁在主张什么、朝哪个方向。句式用 The article argues that + 完整从句,不要用名词短语堆砌。 2. 第 2 句只负责支撑:证据是什么类型(调查、案例、数据、历史对比),关键数字最多给一个。给两个以上就超词。 3. 第 3 句只负责结论或影响:用 The implication is that… 或 The author concludes that…,不加自己的评价、不预测、不提建议。 4. 时态分两层:文章的主张和普遍结论用一般现在时;文中提到的具体历史事件用一般过去时。混用会让读者分不清哪句是作者的话。 5. 句间靠指代衔接,不靠连接词堆叠:this claim、these findings、that gap。三句话里出现两个 however,说明你在写议论文,不是摘要。 第 2 句的证据类型只有四种,写之前先归类:调查或数据(a survey of 12,000 employees)、案例(two cities, one factory)、历史对比(before and after 2019)、专家或机构说法(the central bank estimates)。归不了类,说明你其实没记住文章用了什么证据,回去看 Day 1 划的主题句。 **正误对比** | 错误写法 | 问题 | 改法 | | --- | --- | --- | | The article is about remote work and productivity. | 只报话题,没有主张 | The article argues that remote work raises output only when targets are clear. | | Remote work has become common. Many firms allow it. Some studies disagree. | 三句平行,没有层次 | 三句分别承担主旨、证据、结论 | | The author shows that output rose, so companies should cut office space. | 最后半句是读者自己的建议 | The author shows that output rose; the implication is that measurement matters. | **三个常见错误** - 第 2 句开始讲背景,把证据写成铺垫。摘要没有背景段,第 2 句必须直接给证据。 - 第 3 句重复第 1 句,只换了几个词。判断方法:把两句并排读,如果信息量相同,删掉第 3 句重写。 - 一句里塞两个 that 从句加一个 which 从句,读完不知道重点。摘要句以清楚为先,宁可短。 ## 四、句型与范例 | 功能 | 句型 | 例句 | 中文 | | --- | --- | --- | --- | | 主旨 | The article argues that … | The article argues that fees have hurt small shops. | 文章主张收费伤害了小商铺。 | | 主旨(更强) | Its central claim is that … | Its central claim is that the model cannot scale. | 其核心主张是这个模式无法规模化。 | | 支撑 | It supports this by … | It supports this by comparing two cities. | 它通过比较两座城市来支撑。 | | 支撑(分词) | Drawing on …, the author shows that … | Drawing on survey data, the author shows that hours fell. | 借助调查数据,作者表明工时下降了。 | | 证据来源 | The evidence comes from … | The evidence comes from a three-year trial. | 证据来自一项为期三年的试验。 | | 结论 | The author concludes that … | The author concludes that the rule should stay. | 作者得出结论:规则应保留。 | | 影响 | The implication is that … | The implication is that firms must retrain staff. | 这意味着企业必须重新培训员工。 | | 收束 | In short, … | In short, the gains are real but uneven. | 简言之,收益真实但不均衡。 | ## 五、练习 **A 词汇与搭配(8 题)** 1. The article ______ that the ban has done more harm than good.(argue) 2. The claim is ______ by data from 40 hospitals.(support) 3. The author ______ on interviews with twelve managers.(draw) 4. The trial ______ that the new schedule cut errors.(demonstrate) 5. The report ______ that spending will level off next year.(conclude) 6. The ______ is that smaller suppliers will disappear.(implication) 7. 翻译:总体而言,证据指向同一个方向。→ 用 overall + point to 8. 翻译:简言之,这项政策帮的人少、代价大。→ 用 in short **B 改写(6 题)** 把下面六句改成合格的摘要句,或指出它属于三句中的哪一句、错在哪。 1. The article is about how cities deal with traffic. 2. The author says "the hours saved were often spent on extra meetings" and this is a quote. 3. Many people work from home now. It is very popular. 4. Because output rose, companies should close their offices next year. 5. The article argues that the tax failed. The article argues that the tax did not work. 6. The evidence comes from a survey. **C 输出(1 题)** 给你本周的文章写 100 词以内的摘要初稿,严格三句。每句后面用括号标词数,三句合计不超过 100。至少用出 2 个本周核心词。 ## 六、参考答案 **A** 1. argues(主语单数,一般现在时;后接 that 从句) 2. supported(be supported by,被动) 3. draws(draw on 是固定短语,主语单数加 s) 4. demonstrated(已完成的试验,过去时) 5. concludes(报告结论,现在时) 6. implication(the implication is that,固定句式) 7. **Overall**, the evidence **points to** the same conclusion.(overall 句首加逗号;point to 后接名词) 8. **In short**, the policy helps few people and costs a lot.(in short 后加逗号,句子要短) **B** 1. 错在只报话题。改:The article argues that cities have misread the cause of congestion. 2. 错在照抄原句。改:The author shows that saved commuting time was largely absorbed by meetings. 3. 两句都在说背景,没有主张。改:The article argues that remote work is less popular with managers than with staff. 4. 最后半句是自己的建议,超出摘要范围。改:The author shows that output rose; the implication is that measurement, not location, drives results. 5. 两句同义重复。改:The article argues that the tax failed, and supports this with three years of revenue data. 6. 证据太笼统,没有类型和规模。改:The evidence comes from a survey of 12,000 employees in four countries. **C 参考范文(74 词,三句)** The article argues that remote work saves less time than promised, since hours freed from commuting are often absorbed by extra meetings.(22)Drawing on a survey of 12,000 employees, the author shows that output rose only where managers set explicit targets, and attributes the gap to how work is measured rather than where it happens.(33)The implication is that, absent a consensus on measurement, firms should judge staff by results rather than hours online.(19) 初稿最常见的毛病是「三句合一」:第 1 句就把主旨、证据、结论全说了,后两句只能重复。出现这种情况,不要硬拆句子,回提纲重排——把第 1 句里的证据剪下来放进第 2 句,把结论剪下来放进第 3 句。剪完如果第 1 句只剩十几个词,那才是它该有的长度。 评分要点:三句是否分别对应主旨、证据、影响;是否点名证据类型;是否出现自己的建议或预测;本周词是否搭配正确(尤其 attribute … to)。词数按空格计,连字符算一个词。如果三句合计超过 110 词,问题通常出在第 2 句——证据讲成了故事,删掉过程,只留证据类型和一个数字。 ## 七、今日输出与自查 | 检查项 | 标准 | 完成 | | --- | --- | --- | | 结构 | 严格三句,每句只承担一层功能 | □ | | 词数 | 合计 ≤ 100 词,每句后标了词数 | □ | | 原创 | 与原文没有连续 5 词重合(写完自己比对一遍) | □ | | 用词 | 至少 2 个本周核心词,介词搭配正确 | □ | | 存档 | 存为 Day04_摘要初稿.md,保留提纲三行,Day 5 在此基础上改 | □ | 今天不追求好句子,追求「三句站得住」。如果第 2 句写不出证据,先写「证据类型 + 一个数字」占位,明天再润色;如果第 3 句写成了建议,划掉,改成 The implication is that 加文章自己的结论。 ## 八、明日预告 Day 5 只做一件事:把初稿里的「抄来的痕迹」清理干净。三个动作——删照抄的短语、换同义表达、补必要的连接。明天会给你一张逐词比对的方法,教你怎么在 100 词里既保留原意,又不留下连续五个相同的词。今天写完先别急着改,把初稿放一晚,明天带着新眼睛再看,能发现更多照抄的痕迹。