# 英语高级 · 第 8 周 Day 2|引用三型与改写三换 > 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** 今天解决的是"借了别人的话却说不清来源"的问题。直接引用、转述、综合看着是三件事,其实是同一条链上的三种距离:引语最近,转述中距,综合最远——你也可以谁都不靠,自己下判断。链断了,文章就会出现两种病:要么整段是引语拼接,读者不知道你想说什么;要么通篇是自己的话,却有一句悄悄来自别人。今天写二十个引用例、改十句,标准很硬:改写后与原文零处连续五词重合。这是 Day 5 synthesis 的前置工序,也决定你后面每一段的可信度。 ## 一、今日任务 | 环节 | 时长 | 具体动作 | 产出 | | --- | --- | --- | --- | | 1. 定判据 | 15 min | 给三种引用各写三条"非用不可"的条件:什么时候只能用直引、什么时候必须转述、什么时候才轮到综合 | 三型判据表,每型三条 | | 2. 三型各五例 | 25 min | 取同一篇材料里的同一组观点,分别写成五条直引、五条转述、五条综合;综合至少跨两个来源 | 十五例,逐条标页码或段落 | | 3. 补五例混合 | 10 min | 写五例"引语嵌在转述句里"的混合式,引语不超过十二词 | 五例混合式,共二十例 | | 4. 十句改写 | 20 min | 每句至少用一种换法:换词、换句法、换信息顺序;术语与数字不动 | 十句改写,附所用换法 | | 5. 并排查重 | 10 min | 原文与改写逐句并排,按五词窗口机械比对,标出所有重合处并再改一次 | 十句零重合记录 | 第 2 步的顺序很重要:先写直引,再写转述,最后写综合。因为直引逼你看清原句的措辞,转述逼你拆掉措辞只留意思,综合才可能把两个来源放进同一句而不打架。 ## 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 | | --- | --- | --- | --- | --- | --- | --- | | paraphrase | /ˈpærəfreɪz/(≈派若夫雷兹) | v. / n. | 转述,改写 | paraphrase a passage; a close paraphrase | A paraphrase has to survive being read next to the original. 转述必须经得起与原文并排读。 | 学术 | | verbatim | /vɜːˈbeɪtɪm/(≈韦贝提姆) | adj. / adv. | 逐字的;逐字地 | quote verbatim; verbatim overlap | Anything reproduced verbatim needs quotation marks and a page number. 任何逐字复制的内容都需要引号和页码。 | 学术 | | attribution | /ˌætrɪˈbjuːʃn/(≈艾特里布尤希恩) | n. | 归属,注明出处 | clear attribution; attribution of an idea to X | Attribution is not decoration: it tells the reader whose claim is being tested. 出处标注不是装饰,它告诉读者正在检验谁的主张。 | 学术 | | cite | /saɪt/(≈赛特) | v. | 引用,援引 | cite a study / a figure; as cited in | Cite the version you actually read, not the one it cites. 引用你真正读过的版本,而不是它转引的那一个。 | 学术 | | patchwriting | /ˈpætʃˌraɪtɪŋ/(≈派奇赖提英) | n. | 拼贴式改写(保留原句骨架、替换近义词) | patchwriting from a source | Patchwriting passes a checker and fails a reader. 拼贴式改写骗得过查重,骗不过读者。 | 学术 | | plagiarise | /ˈpleɪdʒəraɪz/(≈普莱哲赖兹) | v. | 抄袭 | plagiarise a passage; avoid plagiarising | Reusing a paragraph's structure without credit is plagiarising even when every word differs. 即使每个词都不同,未加注明而沿用段落结构也算抄袭。 | 学术 | | synthesise | /ˈsɪnθəsaɪz/(≈西恩斯厄赛兹) | v. | 综合(多来源) | synthesise findings across studies | The review synthesises five studies without flattening their disagreements. 这篇综述综合了五项研究,又没有抹平它们的分歧。 | 学术 | | integrate | /ˈɪntɪɡreɪt/(≈伊恩提格雷特) | v. | 整合,使引语融入句子 | integrate a quotation into a sentence | A quotation dropped in whole rarely fits the grammar of the sentence around it. 整句塞进去的引语,很少能与上下文的语法咬合。 | 学术 | | ellipsis | /ɪˈlɪpsɪs/(≈伊利普西斯) | n. | 省略号(表示删去) | mark an omission with an ellipsis | Use an ellipsis only where the omission leaves the sense intact. 只在删去部分不改变原意时使用省略号。 | 学术 | | précis | /ˈpreɪsiː/(≈普雷西·长) | n. | 提要,摘要 | write a précis of | A précis keeps the order of the argument and drops its examples. 提要保留论证顺序,删去例子。 | 学术 | 这十个词可以按"三种距离"排开。最近的一端是 verbatim 与 ellipsis,它们管的是"一字不改时怎么标";中间是 paraphrase、précis、patchwriting,三者都是转述,区别只在成败——précis 只留骨架,paraphrase 保留全部信息但换掉语言,patchwriting 是失败的 paraphrase。最远的一端是 synthesise 与 integrate,它们管的是"多个来源怎么进同一句"。attribution、cite、plagiarise 贯穿三端,是同一件事的三个说法:标了就是 cite,没标就是 plagiarise。 四个搭配要点。cite 后面直接跟宾语,不说 cite from;paraphrase 作名词可数,a paraphrase of the passage;plagiarise 是英式拼写,美式为 plagiarize,全文统一即可;ellipsis 的复数是 ellipses,与"椭圆"同形,读上下文判断。 ## 三、语言点:改写的三层操作与"换词不换骨" 改写失败的典型症状是:每个词都换了,句子读起来还是原文。原因是你只动了第一层。 **规则** 1. **先合上原文再写。** 看着原文改写,眼睛会不断回到原句的语序,手也跟着走。读完、标出三个关键词、合上材料,用自己的话讲一遍,落笔的是你的句子。 2. **换词只换可换的词。** 术语、数字、专名、法条编号一律不动;可动的是普通动词、形容词与名词化结构。把 important 换成 crucial 不算改写,把 plays a crucial role in 换成 drives 才算。 3. **换句法至少用一种。** 从句与名词短语互换、主动与被动互换、拆句与合句互换。三种里用一种,骨架就变了;一种都不用,句子只是换了外套。 4. **换信息顺序,但不换因果方向。** 原文"因为样本小,所以不能推广"可以写成"结论止于城市边界,因为样本只来自一座城市";把因果倒过来就是另一句话,不是改写。 5. **改完再补标注,不边改边标。** 先标注会让你盯着"格式对不对",忘了"意思准不准";改完并排核对,再决定这条该给页码还是只给作者。 **强弱对比** | 原句 | 拼贴(不合格) | 合格改写 | 差别在哪 | | --- | --- | --- | --- | | Because the sample was drawn from a single city, the findings cannot be generalised to rural areas, where referral patterns differ sharply. | Since the sample was taken from one city, the results cannot be generalised to rural areas, where referral patterns are very different. | The study draws its evidence from one city. Rural referral patterns differ, so its conclusions stop at the urban boundary. | 拼贴换了 because、drawn、sharply,骨架与五词串仍在;合格版拆成两句、换了主语、把结论后置 | | Between 2010 and 2019, the share of households renting rose from 31% to 42%. | From 2010 to 2019, the proportion of households renting increased from 31% to 42%. | Renting became the tenure of 42% of households by 2019, up from 31% nine years earlier. | 数字不能动,能动的是句式:前者仍是"时间状语加上升",后者把结果提到句首 | | The report states that "the costs were underestimated by a factor of three", which shows incompetence. | The report says costs were underestimated threefold, showing incompetence. | The report puts the underestimate at "a factor of three" — serious, but not by itself evidence of incompetence. | 第二句既没标引号又把结论塞给作者;第三句保留原话、把结论收回 | **常见误用** 1. **同义词汤。** 逐词替换近义词,句子长度与语序原封不动。查重可能饶过你,读者不会:这种句子读起来像翻译软件,且常常破坏搭配,比如把 carry out a study 换成 execute a study。 2. **改写时顺手加评价。** 原文只说样本来自一座城市,改写里冒出"样本严重不足"。这是把自己的判断记在别人账上,比抄袭更难解释。 3. **省略号改变原意。** 原文"虽有代价,仍然可行",删掉转折分句后只剩"可行"。删的部分只要影响结论方向,就必须补出,不能用省略号盖过去。 **用与不用的分寸** 改写适用于"你要的是意思,不是措辞"的时候,占了学术写作的绝大多数。直引只在三种情况下用:原话本身是证据(当事人陈述、法条、合同条款)、措辞本身有争议(对方用了一个可疑的词)、你要逐字分析某句的逻辑。综合只在你已经读完两个以上来源、并且能说出它们的关系时用;读了一个来源就写"研究表明",那不是综合,是把单例包装成共识。三者用错位置,读者立刻能感觉到:该直引的地方转述,像在躲;该转述的地方直引,像在凑字数。 ## 四、句型与范例 | 句式 | 功能与适用场景 | 中文 | | --- | --- | --- | | As X puts it, "…" | 直引,原话有力且短;引语不超过二十五词 | 正如 X 所说,"……" | | X argues that… | 转述主张,最中性的归因 | X 主张…… | | X's study of 1,200 households found that… | 转述带方法,用来建立可信度 | X 对一千二百户家庭的研究发现…… | | Where X stresses cost, Y stresses timing. | 综合,对比结构最省字 | X 强调成本,Y 强调时机 | | Both studies converge on one point: … | 综合,先给共识 | 两项研究在一点上趋同:…… | | The two accounts diverge over… | 综合,先给分歧 | 两种说法在……上分歧 | | X's claim that… has since been qualified by Y, who notes… | 综合加时间维度,适合文献有推进时 | X 关于……的主张后来被 Y 限定,Y 指出…… | | Read together, these findings suggest… | 你自己的判断,不归因任何一方 | 合起来看,这些发现表明…… | | X goes further than Y, arguing that… | 综合,标出强度差 | X 比 Y 更进一步,主张…… | | The evidence is mixed: X finds…, whereas Y finds… | 综合,中立呈现冲突 | 证据并不一致:X 发现……,而 Y 发现…… | **示范段(约一百五十词,三型混用)** > Whether minimum wages cost jobs depends on how the question is asked. Neumark and Wascher argue that the least skilled lose hours when the floor rises, while Dube and colleagues, reviewing the same period, find "no measurable employment effect" in states that raised their minimum. Read together, the two bodies of work suggest that the answer is local: where wages are already close to the median, the effect is small; where they are far below it, employers adjust. That reading is not a compromise between the studies but a claim about their scope — and it is testable, since it predicts where the disagreement should disappear. 批注三条。第一,两位作者用 argue,第三位用 find,动词不同是因为一个给判断、一个给结果,归因强度也随之不同。第二,引语只取八个词,且是原文的关键措辞,其余全部转述,句子因此保持自己的语法。第三,Read together 之后的那句不再归因,它是作者自己的判断,也是整段的落点——没有这一句,前面两个来源只是并列,不构成综合。 ## 五、练习 **A 词汇与搭配(7 题)** 1. 填空:Every figure in the memo should be ______ to a published table.(用 trace 的形容词形式) 2. 改搭配:The author cites from three separate surveys. 3. 选词(patchwriting / précis):A ______ keeps the argument's order but drops its examples. 4. 中译英:这份摘要曲解了研究发现:它测的是相关,不是因果。 5. 词形转换:synthesise → 名词;integrate → 名词。 6. 语域升格:把 A study says remote work is fine. 改成学术语域的一句,须含样本与年份。 7. 造句:用 integrate 写一句,宾语是 a quotation,并说明它融入了什么。 **B 句法(6 题)** 1. 改写(换句法):Although the programme was widely praised, it failed to reach the poorest households. 用名词化开头重写。 2. 改写(换顺序):The scheme was abandoned because its costs tripled within two years. 把结论提前。 3. 判断重合:下面哪一组构成连续五词重合?a. 原文 the findings cannot be generalised to rural areas/改写 these results cannot be applied to rural districts;b. 原文 the share of households renting rose/改写 the share of households renting climbed。 4. 把直引改成转述:As one official said, "we had no idea the backlog would grow this fast". 5. 合并两源为一句:Study A finds the effect is small. Study B finds the effect is small in cities only. 6. 删到二十词以内,保留归因:According to a 2021 review of forty studies conducted in eleven countries, the policy had, on average, no detectable effect on employment. **C 输出(2 题)** 1. 二十例引用表:直引五、转述五、综合五、混合五,每例标来源与页码或段落,综合例须跨两个来源。 2. 三百词短文:就"最低工资是否影响就业"写一段三型混用的文字,至少一个直引、两个转述、一处综合,结尾必须有一句自己的判断。 ## 六、参考答案 **A** 1. traceable。be traceable to 是固定搭配,介词用 to。 2. The author cites three separate surveys. cite 为及物动词,没有 cite from。 3. précis。patchwriting 是失败的改写,不留骨架只留措辞痕迹,与"保留顺序、删去例子"相反。 4. The summary misrepresents what the study found: it measured correlation, not causation. 5. synthesis;integration。 6. A 2022 survey of 3,000 knowledge workers found no decline in self-reported productivity. 关键是给出样本与年份,把 a study 变成可核查的来源。 7. The paragraph integrates a twelve-word quotation into its own syntax, so the sentence still reads as one sentence. 检验标准是:删掉引号,句子是否仍合语法。 **B** 1. The programme's wide acclaim did not prevent it from failing to reach the poorest households. 名词化作了主语,转折由 prevent 承担。 2. The scheme was abandoned: its costs had tripled within two years. 冒号把结论提前,因果方向未变。 3. b。a 中 cannot be generalised 与 cannot be applied 不同,且 rural areas/rural districts 已换;b 的 the share of households renting 五词连排,构成重合。 4. One official later admitted that the growth of the backlog had taken them by surprise. 去掉引号、换掉措辞、保留"意外"这一核心意思。 5. Both studies find the effect small, though B locates it in cities only. 用 though 承接分歧,避免两句并列。 6. A 2021 review of forty studies across eleven countries found no detectable employment effect. 二十词以内,归因与范围都保留。 **C 评分维度** 第 1 题:二十例是否同一材料、同一组观点(同题才看得出三型差别);综合例是否真跨两源;混合例的引语是否 ≤12 词且语法咬合。第 2 题:直引是否只用于措辞本身重要的地方;转述是否零连续五词重合;综合句是否给出关系(趋同、分歧、限定),而不是把两句话用 and 连起来;末句是否是你自己的判断且可被检验。 ## 七、今日输出与自查 | 检查项 | 合格线 | 我的记录 | | --- | --- | --- | | 例数 | 二十例,四型齐全 | | | 来源标注 | 每例可追溯到页码或段落,无"网上资料" | | | 改写重合 | 十句均零处连续五词重合 | | | 三型判据 | 每型三条,写成"什么时候非用不可" | | | 三百词段 | 含一处综合与一句自己的判断 | | 产出文件名:引用二十例_Day02.md、改写十句_Day02.md、三型短段_Day02.md。改写那十句保留原文与改写并排,Day 6 查重时直接拿它当模板。 ## 八、明日预告 Day 3 把镜头拉近到一句话内部:让步与反驳怎么在一句里完成。今天你练的是"别人的话怎么进来",明天练的是"进来之后怎么退一步再压回去"——While it is true that…、this objection overlooks…、even if…、the claim that… is undermined by…,四个句式各写三句,共十二句,每句都要能指出哪半句是退、哪半句是进。写不出"进",说明你还没找到自己真正的理由。