# 英语高级 · 第 8 周 Day 6|改稿三查与朗读一遍 > 系列:语种学习 / 英语 / 高级 · 每日学习|AI生成|2026-10-27|codebuddy **今天学什么** 今天六十分钟只做改稿,而且顺序不能乱:先核引用,再查连续五词,再体检连接词,最后朗读一遍。多数人改稿从"读起来顺不顺"开始,结果把一句准确的转述改成了不准确但顺口的句子,后面两查全部白做。今天要建立的是工序意识——事实先于文字,文字先于衔接,衔接先于语感。改完的稿子是二稿,明天只做口头交付,不再动结构。六十分钟不长,所以今天每一步都有明确的停止条件,不做无限打磨。 ## 一、今日任务 | 环节 | 时长 | 具体动作 | 产出 | | --- | --- | --- | --- | | 1. 核引用 | 20 min | 每处引用回原文,填三栏:原文怎么说、我怎么用、是否支持(只填是或否) | 引用核对表,凡"否"必改或删 | | 2. 查重合 | 15 min | 原文与二稿各断成五词窗口,逐句机械比对,标出所有连续五词重合 | 重合清单,逐处改到零 | | 3. 体检连接 | 15 min | 统计每段段首三词与全文连接词,同类超过两次就换或删 | 连接词统计表 | | 4. 朗读 | 10 min | 全文朗读一遍,只画线不改:卡顿、换气、需要回读的地方 | 朗读标记稿,至少五处 | 今天的时长按计划为六十分钟,是本周唯一的短日。如果状态好,可以把朗读延长到十五分钟读第二遍,但不改验收标准:引用零错误、重合零处、连接词不重复。 ## 二、核心词汇与搭配 | 词/短语 | 音标 | 词性 | 释义 | 高频搭配 | 例句(含译) | 语域 | | --- | --- | --- | --- | --- | --- | --- | | misquote | /ˌmɪsˈkwəʊt/(≈米斯克沃特) | v. | 误引 | misquote a source; be misquoted | Trimming a clause can misquote a source without changing a single word. 删掉一个从句,就可能一字未改地误引来源。 | 学术 | | distortion | /dɪˈstɔːʃn/(≈迪斯陶希恩) | n. | 歪曲 | a distortion of the evidence | The summary is a distortion by omission: the null result disappears. 这份摘要是靠省略造成的歪曲:零结果不见了。 | 学术 | | traceable | /ˈtreɪsəbl/(≈特雷瑟布尔) | adj. | 可追溯的 | traceable to a page | Every quotation should be traceable to a page, not to a search result. 每处引语都应能追溯到页码,而不是搜索结果。 | 学术 | | overstate | /ˌəʊvəˈsteɪt/(≈欧弗斯泰特) | v. | 夸大 | overstate the case; it is hard to overstate | "Proves" overstates what a single trial can show. "证明"夸大了单项试验能做到的事。 | 通用 | | cohesion | /kəʊˈhiːʒn/(≈科希日恩·长) | n. | 衔接,连贯 | improve cohesion; cohesive | Cohesion is carried by pronouns and repeated key terms, not by connectives alone. 衔接靠代词与关键词复现,不只靠连接词。 | 学术 | | redundancy | /rɪˈdʌndənsi/(≈里丹德恩西) | n. | 冗余 | cut redundancy; redundant phrasing | The draft's redundancy sits mostly in paired synonyms: "first and foremost", "each and every". 草稿的冗余多半在成对同义词上。 | 学术 | | overlap | /ˈəʊvəlæp/(≈欧弗莱普·轻) | n. / v. | 重合 | word-for-word overlap; overlap with the source | Five consecutive words of overlap is a working threshold, not a legal one. 连续五词重合是工作阈值,不是法律标准。 | 学术 | | prune | /pruːn/(≈普鲁恩·长) | v. | 删减 | prune a draft; prune back | Prune first, then rewrite: deletion exposes the sentences that were doing no work. 先删再改:删减会让不出力的句子露出来。 | 通用 | | stumble | /ˈstʌmbl/(≈斯塔姆布尔·短) | v. | 卡住,绊一下 | stumble over a phrase | You will stumble exactly where the syntax outruns the thought. 你卡住的地方,正是句法跑在思想前面的地方。 | 通用 | | read-through | /ˈriːd θruː/ | n. | 通读(尤指朗读一遍) | a full read-through | One read-through aloud finds more than two silent proofreads. 朗读一遍发现的错误,比两次默读还多。 | 通用 | 这十个词是改稿的判断工具,不是装饰。misquote 与 distortion 的区别在意图:前者可以无意,后者通常指效果;写自评时用 misquote 描述操作失误,用 distortion 描述读者可能得到的结果。traceable 是引用的硬标准,做不到就删。overstate 是动词里的刹车,凡是 prove、always、never 出现,先想它。cohesion 与 redundancy 是一对:前者缺了要补,后者多了要删,两者都靠删读判断。overlap 是量词化的标准,prune 是动作,stumble 与 read-through 是检验手段。 两个搭配要点。overlap 作名词时与 with 连用,overlap with the source;作动词时可直接跟宾语,the two paragraphs overlap。read-through 是可数名词,a read-through、two read-throughs,不写 do a read through 的分词形式。 还有一个语域提醒。自评时用 misquote、overstate 这类描述操作的词,读者会认为你在谈技术;用 distortion、failure 这类描述结果的词,读者会认为你在下判断。改稿记录是给自己看的,越具体越有用:写"第三段的引语早停一个从句",比写"引用有问题"更能指导下一次。 ## 三、语言点:三查的顺序为什么不能颠倒 改稿不是把句子弄顺,是在四个层面上依次排错。顺序颠倒,后面的工作会毁掉前面的结果。 **规则** 1. **先查事实,再查文字。** 引用核对放在第一步,因为改句子会改变转述的重心:把"部分研究"改成"研究",语法更顺,引用却失效了。先改文字,你会在不知不觉中制造新的误引。 2. **重合检查必须机械。** 把原文与自己的句子都断成五词窗口,逐一对照。凭印象判断"这句是我自己写的",几乎一定会漏,因为熟悉的措辞读起来最像自己的。 3. **连接词体检看段首三词。** 逐段抄下前三词,同类开头超过两次就换或删。全文 however 出现四次以上,说明你在用连接词替代论证。 4. **朗读只做标记,不做修改。** 边读边改会让你停下来打磨一句,丢掉整篇的节奏感;画线,读完统一处理。卡顿处通常有两个原因:句子超过三十五词,或逻辑在这里断了。 5. **改稿以删为主。** 删掉一句,段落通常更清楚;加一句,就要重新核对来源。二稿阶段新增内容,是引用错误的主要来源。 **强弱对比** | 改前 | 改后 | 差别在哪 | | --- | --- | --- | | Moreover, furthermore, the study also additionally finds that costs rose. | The study also finds that costs rose. | 三个连接词做同一件事,删两个,句子反而更有力 | | This proves that remote work damages careers. | This is consistent with remote work damaging careers. | prove 要求排他性,单一研究给不出;consistent with 与证据量匹配 | | The survey, which was conducted in 2019 and covered 4,000 nurses across eleven hospitals and is therefore one of the largest of its kind, found… | A 2019 survey of 4,000 nurses in eleven hospitals found… | 前者把方法写成了定语从句,读者到 found 已经忘了主语;后者把方法压成前置定语 | **常见误用** 1. **改稿时加新论点。** 新论点的来源没核对过,等于在最后一步引入风险;二稿只做减法与替换,不加主张。 2. **用 however 覆盖所有转折。** 让步用 admittedly,对比用 by contrast,因果用 as a result,三者在逻辑上不等价,混用会让读者误判关系。 3. **为降重把术语换成近义词。** 把 cohort 换成 group、把 attrition 换成 loss,重合率下来了,准确性也没了;术语与数字一样,属于不可换的词。 **用与不用的分寸** 这套工序适用于任何要交出去的文本:论文、memo、提案。不适合用于创作型写作,那里语感优先。三查的严格程度可以调:学术写作三查全做,商务写作可省第二步(若无外部来源),口语稿只做第四步。但顺序不能调——先事实后文字这一条,任何文体都成立。 ## 四、句型与范例 | 修订句式 | 修订动作 | 中文 | | --- | --- | --- | | This is consistent with…, though it does not establish… | 降强,把 prove 换成匹配证据的说法 | 这与……一致,但不能确立…… | | The study finds X; it does not find Y. | 分句划界,防止读者扩大结论 | 该研究发现 X,并未发现 Y | | Read narrowly, the figure says only that… | 限定解读范围 | 严格解读,这个数字只说明…… | | The same point can be put without the jargon: … | 去术语,检验自己是否真懂 | 同一意思可以不用术语说:…… | | What the paragraph needs is a subject, not another connective. | 改稿判断,用于自评 | 这段需要的是主语,不是又一个连接词 | | Cut the first clause; the sentence gets its verb back. | 删冗,恢复主谓 | 删掉第一个从句,句子就找回动词了 | | The quotation stops one clause too early. | 指出误引的具体形态 | 这句引语早停了一个从句 | | On re-reading, the claim is broader than the evidence. | 二稿总评 | 重读之后,主张比证据更宽 | **改前改后对照(约一百四十词)** 改前:Moreover, it is also important to note that the 2019 survey, which was conducted among 4,000 nurses and is therefore quite representative, proves that remote work damages careers, and this is a very important finding. 改后:A 2019 survey of 4,000 nurses found that promotion rates fell by a fifth for staff under thirty. That is consistent with remote work damaging careers; it does not prove it, since the survey cannot separate remote work from the recession that began the same year. 批注三条。第一,改后把方法压成前置定语,主语回到 the survey,读者第一眼就知道谁做了什么。第二,改后加了限定句 it does not prove it, since…,这不是软化立场,而是把结论与证据对齐;对齐之后,读者反而更愿意相信前半句。第三,改后删掉了开头三个连接词与结尾的评价句 this is a very important finding——评价由事实承担,读者自己会判断它重不重要。 ## 五、练习 **A 词汇与搭配(7 题)** 1. 填空:Every claim in the essay should be ______ to a specific page.(trace 的形容词形式) 2. 改搭配:The two paragraphs overlap with each other. 3. 选词(misquote / distortion):By dropping the final clause, the summary turns a qualified finding into a ______. 4. 中译英:连续五词重合是工作阈值,不是法律标准。 5. 词形转换:cohesion → 形容词;redundancy → 形容词。 6. 语域升格:把 This proves the policy failed. 改成与证据量匹配的一句。 7. 造句:用 prune 写一句,宾语是 a draft,并说明删完之后发生了什么。 **B 句法(6 题)** 1. 删冗:Although it is true that costs rose, but the increase was modest and, moreover, entirely expected. 2. 降强:The data demonstrate that the scheme cannot work. 3. 恢复主谓:The report, which was commissioned in 2020 by the ministry and covered eleven regions over three years, recommends consolidation. 4. 换连接:把两个 however 中的一个换成表示让步或对比的连接语。 5. 加限定:The survey shows that remote work reduces promotion. 6. 指出误引:原文说 "the effect was modest in the first year",改写写成 "the effect was modest"。问题在哪? **C 输出(2 题)** 1. 二稿:在初稿上改,保留修改痕迹;引用零错误,连续五词零重合,连接词同类不超过两次。 2. 朗读记录表:至少五处卡顿,每处写原因(句长、逻辑断点、生词)与处理方式。 ## 六、参考答案 **A** 1. traceable。be traceable to 为固定搭配。 2. The two paragraphs overlap. overlap 作动词及物,with each other 冗余。 3. distortion。此处描述效果,不涉及是否故意;misquote 指误引这一动作。 4. Five consecutive words of overlap is a working threshold, not a legal one. 5. cohesive;redundant。 6. The data are consistent with the policy having failed. 或 On this evidence, the policy appears to have failed. 7. Prune the draft before rewriting; what remains is usually the argument. 删减之后,剩下的才是论证。 **B** 1. Costs rose, but the increase was modest and entirely expected. 删 although、but 之一与 moreover。 2. The data suggest that the scheme may not work. demonstrate 与 cannot 都过强。 3. A 2020 ministry report covering eleven regions over three years recommends consolidation. 方法压成前置定语,主语与动词回到原位。 4. 示例:The first objection is technical. By contrast, the second is political. 用 by contrast 替换第二个 however。 5. The survey shows that remote work is associated with lower promotion rates among junior staff. 加范围与"相关"限定。 6. 删掉了时间限定 in the first year,读者会以为效应一直很小;引用时删任何限定语,都必须确认结论方向未变。 **C 评分维度** 第 1 题:引用核对表是否每处填"是/否",凡"否"是否改或删;重合清单是否归零;连接词统计是否同类 ≤2;修改痕迹是否看得出删多于加。第 2 题:五处卡顿是否各写原因;原因是否落在句长、逻辑、生词三类;处理方式是否具体到"拆句、加主语、换词"。 ## 七、今日输出与自查 | 检查项 | 合格线 | 我的记录 | | --- | --- | --- | | 引用 | 零错误,每处支持所标注的判断 | | | 重合 | 连续五词零重合 | | | 连接 | 同类连接词全文不超过两次 | | | 朗读 | 至少标出五处,且写明原因 | | | 改动方向 | 删多于加,未新增论点 | | 产出文件名:synthesis二稿_Day06.md、朗读记录_Day06.md。二稿保留修订模式,明天终稿只做口头对应的时间标记,不再改论证。 ## 八、明日预告 Day 7 是本周唯一的高压日:八分钟辩论录音,立论三分钟、驳论三分钟、结辩两分钟,外加提交四百词终稿。今天查出的每一处引用错误,明天都会在驳论里变成对方的弹药;今天朗读标出的每一处卡顿,明天都会变成超时的十秒。把二稿放在手边,明天上午先写三个论点和各自的证据,再动嘴。