# 英语高级 · 第 6 周 Day 1|深度报道论证图五节点 > 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** 今天解决一个很具体的困境:1200 词的深度报道,每一句都读懂了,合上却讲不出作者凭什么这么说。原因是你读到的是句子,不是论证。今天把一篇 FT 深度报道拆成五类节点——主张、证据、隐含前提、反驳、结论,再用箭头标出谁支撑谁、谁削弱谁,逼自己看见句子之间的承重关系。这是本周所有任务的地基:第二天的立场标记、第三天的归属改写、第五天的 250 词立场分析,都要落在这张图上。学完你手里应有一张节点不少于五个的论证图,并能指着它用 90 秒复述全文。 ## 一、今日任务 | 环节 | 时长 | 具体动作 | 产出 | | --- | --- | --- | --- | | 1. 预读定调 | 10 min | 只读标题、副标题、首段、末段与全部小标题,用一句话写下"作者想让我相信什么" | 一句预判主张 | | 2. 限时通读 | 25 min | 计时读完全文,不查词、不回读;页边只标三类记号:→ 因果、≠ 转折、"" 直接引语 | 带三类记号的原文 | | 3. 抽节点 | 20 min | 另取一页,把长句压成短语,填进五栏:主张/证据 1—3/隐含前提/反驳/结论 | 五栏节点清单 | | 4. 连箭头 | 20 min | 画图;实线箭头写 supports,虚线箭头写 weakens;圈出你认为最弱的一条连线 | 论证图 1 张(节点 ≥5) | | 5. 复述校验 | 10 min | 合上文章,只看图讲 90 秒;讲不顺的地方就是图断掉的地方,回去补 | 1 处断点记录 + 200 词图注 | 时间紧时按 README 的降级原则处理:只读 400 词,画出主张、两条证据、一个前提共四个节点,也算不断链;但本周的验收线仍是完整 85 分钟版。 ## 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 | | --- | --- | --- | --- | --- | --- | --- | | contention | /kənˈtenʃn/(≈克恩泰恩希恩) | n. | 论点;争执 | the central contention that…; support the contention | The report's central contention is that the shortfall is structural rather than cyclical. 报告的核心论点是:缺口是结构性的,而非周期性的。 | 学术 | | purport | /pəˈpɔːt/(≈珀保特) | v. | 声称,自称(常含怀疑) | purport to show / to do | The figures purport to show a causal link; they establish only correlation. 这些数字自称显示因果,实际只证明了相关。 | 媒体 | | substantiate | /səbˈstænʃieɪt/(≈瑟布斯坦希诶特) | v. | 证实,提供证据支持 | substantiate a claim / an allegation | None of the leaked documents substantiates the allegation of collusion. 泄露文件没有一份能证实串通的指控。 | 学术 | | premise | /ˈpremɪs/(≈普雷米斯·短) | n. | 前提 | rest on a false premise; an unstated premise | The argument rests on the premise that demand recovers within a year. 该论证建立在需求一年内回升这一前提上。 | 学术 | | infer | /ɪnˈfɜː(r)/(≈伊恩费尔) | v. | 推断 | infer X from Y; invite readers to infer | Readers are invited to infer that the two events were connected. 读者被引导去推断这两件事有关联。 | 学术 | | bolster | /ˈbəʊlstə(r)/(≈博尔斯特尔) | v. | 支持,增强 | bolster a case / an argument | Two independent surveys bolster the paper's case. 两项独立调查支持了该文的论点。 | 媒体 | | undercut | /ˌʌndəˈkʌt/(≈阿恩德卡特) | v. | 削弱,拆台 | undercut an argument / a claim | Its own figures undercut the argument it advances. 它自己引用的数字削弱了它提出的论点。 | 媒体 | | implicit | /ɪmˈplɪsɪt/(≈伊姆普利西特) | adj. | 隐含的,未言明的 | an implicit assumption / premise | The implicit assumption is that households will keep borrowing. 隐含的假设是家庭会继续借贷。 | 学术 | | tenuous | /ˈtenjuəs/ | adj. | 牵强的,薄弱的 | a tenuous link / connection | The link between the two datasets is tenuous at best. 两组数据之间的联系充其量也很牵强。 | 媒体 | | warrant | /ˈwɒrənt/(≈沃若恩特) | v. | 使……有理由;n. 依据 | warrant the conclusion; without warrant | The evidence does not warrant so confident a conclusion. 证据不足以支持如此笃定的结论。 | 学术 | 这十个词分两组。前五个是"拆解用":contention 指主张本身,premise 指它脚下的地,infer 指从证据走到结论的那一步,substantiate 问的是"有没有东西撑着",purport 一出现就提醒你作者在保持距离。后五个是"评价用":bolster 和 undercut 正好是一对方向相反的箭头词,implicit 专用于那些不写在纸上的前提,tenuous 是你能给一条连线的最轻的否定,warrant 则是学术写作里最经济的"够不够"。 两个搭配陷阱先记住:bolster 后面直接跟宾语,没有 bolster up(bolster up 是口语里的冗余);premise 作"前提"讲读 /ˈpremɪs/,作"房屋、场所"讲读 /ˈpremaɪz/ 且多用复数 premises,同一拼写两个音,读错会在正式场合露怯。 ## 三、语言点:从句子的形式判断它在论证的哪一层 深度报道的句子不是平铺的,五种功能各有各的语法外衣。学会看外衣,你就不必读第二遍才找到主张。 **规则** 1. **主张句偏爱现在时与无施事的抽象主语。** "The case for subsidies rests on a misreading of the data."——主语是 case,动词是现在时 rests,没有人出现。作者把判断说成事物的属性,读起来像事实。 2. **证据句偏爱过去时、具体施事与可核查标记。** 时间(in the year to June)、机构(the OECD)、数字(a four-to-one ratio)、直接引语,四样里至少出现两样。缺了这些,所谓证据只是另一个主张。 3. **隐含前提不长成句子,长成修饰成分。** 它藏在定语从句、同位语和形容词里:"a £40bn programme that has yet to break ground"——"尚未开工"是修饰语,但整段论证靠它成立。删掉这个从句,句子依然通顺,论证却塌了。 4. **结论句带情态与回指。** This suggests that…/It follows that…/The implication is that…。情态越弱(may, could),作者越不愿为那一步负责;一旦出现 will、cannot,说明他已经把推测当成了断定。 5. **转折与限定是削弱信号。** yet、although、only、fail to、remains、at best,这六个词附近通常就是全文最脆弱的地方,也是你画虚线箭头的位置。 **强弱对比** | 原句(弱) | 改写(强) | 差别在哪 | | --- | --- | --- | | The policy has been a failure. | Output has fallen for six quarters, against a target of flat growth. | 前者是定性,后者给出可核查的对照;前者无法反驳,后者可以被检验 | | It is clear that the model is finished. | The evidence suggests the model may have run its course. | 删掉 it is clear,加上 may,作者从"宣布"退回"判断",可信度反而上升 | | Costs rose, so the scheme is unworkable. | Costs rose; whether that makes the scheme unworkable depends on what the alternatives cost. | 原句把两个动作压成一个 so,中间那步(成本高到不可行)被藏起来了 | **常见误用** 1. **把形容词当证据。** "The scheme is bloated and the figures are damning."——damning 是评价,不是数据。改写时必须问:哪一张表、哪一年、跟什么比。 2. **把作者的让步读成立场。** "Critics have a point, but…" 的重点永远在 but 之后;只抄前半句,就会把作者写成反方。 3. **自己补前提时补得太狠。** 作者没说"三年足够",你偏写成"作者认为任何政策三年必见效",那是稻草人。补前提的标准是:不补它,结论推不出来;补了它,作者大概率会承认。 **用与不用的分寸** 这套判断只在"分析他人文本"时用。写自己的文章时不要反过来模仿第 1 条——用抽象主语把判断伪装成事实,是社论腔最容易露馅的地方,学术写作里会被要求补出施事(who argues, on what basis)。第 3 条也一样:分析时你要把定语里的前提挖出来,写作时你则该把它写成一句话,让读者有机会反驳。至于第 5 条,only 与 at best 用一次是精确,一段里用三次就显得刻薄,读者会开始怀疑你在选材。 ## 四、句型与范例 | 句式 | 功能 | 中文 | | --- | --- | --- | | The article's central contention is that… | 一句话锁定主张,把"作者认为"换成文本属性 | 该文的核心论点是…… | | The case rests on the premise that… | 指名论证脚下那块地 | 该论点建立在……这一前提上 | | Evidence for this comes chiefly from… | 交代证据来源,避免逐段复述 | 其证据主要来自…… | | The piece moves from X to Y without warrant. | 指出推论跳跃,是最经济的批评句式 | 文章从 X 直接跳到 Y,缺少依据 | | That inference assumes, but never states, that… | 显性化隐含前提 | 该推论假定却从未言明…… | | The weakest link is not the data but the step from… to… | 把最弱环节定位到"那一步"而非材料 | 最弱的一环不是数据,而是从……到……那一步 | | The piece concedes that…, but treats it as… | 处理让步,防止误读立场 | 文章承认……,却把它当作…… | | Read against the final paragraph, the earlier claim looks… | 用结尾反照开头,暴露前后强度不一致 | 对照末段来看,前面的说法显得…… | | On this evidence, the conclusion is at best provisional. | 给整体下判断,留有余地 | 就这些证据而言,结论最多只是暂定的 | **范文:一段 150 词的图注** > The article's central contention is that Europe's chip subsidies will fail on their own terms. The evidence is threefold: a comparison of announced capacity with capacity actually built, interviews with two former officials, and a budget line showing that running costs were never costed. The piece then moves to a conclusion broader than that evidence supports: from the fact that output has not yet risen, it infers that industrial policy cannot work in this sector. That inference assumes, but never states, that three years is long enough for a policy of this kind to show results. The weakest link is therefore not the data but the step from "not yet" to "never". Readers who accept the premise will find the piece persuasive; those who do not will find it merely impatient. 批注四条: 1. 首句用 central contention,而不是 the author thinks,把主张变成可检验的文本对象;这一步决定了整段图注的语域。 2. 证据用 threefold 加冒号列举,三类各一个短语,不复述段落——图注的任务是标结构,不是再讲一遍故事。 3. "a conclusion broader than that evidence supports" 一句同时完成描述与评价:先说它做了什么,再说它做过了头,比直接写 I disagree 有力。 4. 末句把判断交回读者,用 readers who… 两种人对照,作者(你)的责任边界清楚:你指出前提,不替读者下结论。 ## 五、练习 **A 词汇与搭配(7 题)** 1. 填空:None of the cited documents ______ the claim that the scheme had been fully costed.(用 substantiate 的适当形式) 2. 改搭配:The argument is bolstered up by two independent surveys. 3. 选词(tenuous / implicit):The connection between the two figures is ______ at best. 4. 中译英:这一结论建立在"需求会回升"这一未言明的前提上。 5. 词形转换:implicit → 副词;substantiate → 名词。 6. 语域升格:把 The article says really strongly that the plan will fail. 改成媒体或学术语域的一句。 7. 造句:用 undercut 写一句,主语必须是"证据、数字、文件"一类,不能是人。 **B 句法(6 题)** 1. 层级判断:下列三句分别是主张、证据还是结论? a. Output has fallen for six consecutive quarters. b. The subsidy model has run its course. c. It follows that a different instrument is needed. 2. 显性化前提:The £40bn programme, which has yet to break ground, will create 20,000 jobs. 写出它依赖但没说出口的前提(一条即可)。 3. 压缩到 20 词以内,保留转折:Although the government has repeatedly said the target remains achievable, and although independent analysts have questioned the assumptions on which it rests, no revised forecast has been published. 4. 降强:The data prove that the policy failed. 改成与证据量匹配的句子。 5. 合并成一句"主张 + 证据",用冒号或 with 结构。 6. 给结论加距离:It is clear that the model is finished. **C 输出(2 题)** 1. 论证图 1 张:手绘拍照或软件绘制均可,节点 ≥5(主张 1、证据 1—3、隐含前提 1、结论 1),每条箭头标注 supports 或 weakens,圈出最弱连线。 2. 图注 200 词:用本节四类句式写,必须包含"最弱环节"的明确判断,不出现 I think。 ## 六、参考答案 **A** 1. substantiates(主语 None 为单数,现在时;也可用 has substantiated)。 2. The argument is bolstered by two independent surveys.(bolster 为及物动词,up 冗余) 3. tenuous。implicit 修饰"前提、假设",不能修饰"联系"。 4. The conclusion rests on an unstated premise: that demand will recover. 5. implicitly;substantiation。 6. The article contends, in unusually blunt terms, that the plan will fail. / The article's verdict on the plan is unambiguous: it will fail. 7. 示例:Its own budget figures undercut the case for a third round of subsidies.(主语是 figures;若写成 The author undercuts…,指向的是人而非论证,语域就偏了) **B** 1. a 证据(过去时、可核查、无评价);b 主张(现在时、抽象主语、定性);c 结论(回指 + 情态)。 2. 示例:项目能在承诺期内开工,且每单位投资创造的岗位数不低于历史均值。补一条即可,关键是不补它 20,000 这个数字推不出来。 3. The target stands, but the assumptions behind it have been questioned and no new forecast published.(19 词,转折保留,两个 although 合并为一个 but) 4. The data are consistent with the policy having failed. / On this evidence, the policy appears to have failed.(prove 要求排他性,而单一数据做不到) 5. 示例:The case rests on a single comparison: capacity announced has outpaced capacity built by four to one. 6. The available evidence suggests the model may have run its course.(或 The model looks, on this evidence, to be finished.) **C 评分维度**(不自评"写得好不好",逐项打勾) - 节点是否齐全,且"隐含前提"是作者没说、论证必需的一条,不是文中句子的改写; - 箭头是否有方向词,是否至少有一条 weakens; - 最弱环节是否定位到"推论那一步",而不是笼统说"证据不足"; - 图注是否 ≤200 词,是否出现 I think,是否有一句整体判断。 ## 七、今日输出与自查 | 检查项 | 合格线 | 我的记录 | | --- | --- | --- | | 节点数 | ≥5,且五类各至少 1 个 | | | 箭头标注 | 全部写 supports / weakens,无"相关""有关" | | | 隐含前提 | 删掉它结论推不出;补上它作者会承认 | | | 最弱环节 | 指出的是"从 X 到 Y 那一步",不是材料本身 | | | 复述 | 90 秒内只看图讲完,不回原文找句子 | | 产出文件名:论证图_Day01(照片或 PDF)+ 图注_Day01.md(≤200 词)。录音不需要,但复述建议手机录一遍,回听时数一数自己卡了几次。 ## 八、明日预告 Day 2 换一种文本:1000 词的 The Economist op-ed。昨天你画的是骨架,今天要标态度——同一个"削减补贴"的主张,用 prudent 还是 reckless,用 may 还是 will,用 point out 还是 claim,读者接收到的立场强度完全不同。明天要在 80 分钟内标出 20 处立场语言,其中至少 15 处能说清"凭什么判断作者倾向这一边"。带上今天的图,明天你会发现:最弱的那条连线,往往正是形容词最密集的地方。