# 英语高级 · 第 7 周 Day 4|数据演讲听记复述 > 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** 书面数据句可以反复核对,口语只给你一次机会。今天要解决的表达问题是"数字的顺序":多数人复述数据时按听到的顺序报数,听众记不住也听不出重点;成熟的讲法是先给结论,再给两三个带锚点的数字,最后补一句限定。同时练近似语的分层——about、roughly、in the order of、north of 的模糊程度不同,用错会让一个精确的数字听起来像猜测,或者让一个估计听起来像定论。这是本周唯一一次纯输入日,也是第 7 天限时口头分析(如果需要)的预演。学完你能交出一张五数字清单,并用 2 分钟说清每个数字支撑什么观点。 ## 一、今日任务 | 环节 | 时长 | 具体动作 | 产出 | | --- | --- | --- | --- | | 1. 选材与预测 | 10 min | 选一段 10 分钟以上、以数据为主线的讲座或访谈音频(使用你选定的讲座/访谈音频);先看标题写 3 句预测:主讲人会用什么数字证明什么 | 预测 3 句 | | 2. 第一遍无字幕 | 15 min | 只记数字与单位,不记句子;听到比较级、倍数、占比就画圈;没听清的空着,不回退 | 数字草记 | | 3. 第二遍核对 | 15 min | 补全数字,同时记下每个数字的"锚":比什么、哪一年、哪一群人;标出主讲人自己加的限定语 | 数字 + 锚点 | | 4. 整理清单 | 20 min | 按"数字 / 锚点 / 含义 / 支撑哪个论点 / 主讲人的态度"五列填 5 行;删掉那些只是装饰、不支撑论点的数字 | 清单 1 张 | | 5. 复述录音 | 20 min | 2 分钟口头复述,结构固定:一句结论、三个数字、一句限定;不许看稿,只允许看清单上的五个关键词 | 录音 1 条 | | 6. 回听标记 | 5 min | 回听并写 3 个具体问题(不许写"不够流利"):如"第二个数字没给锚点""两处填充词连用" | 3 条问题 | 第一遍听不清数字是正常的。真正要练的是:即使漏掉一个数,你仍然能用其余数字把论点撑住——这要求你在听的时候就判断哪个数字是主证据,哪个只是背景。 ## 二、核心词汇与搭配 | 词 / 短语 | 音标 · 词性 | 释义与高频搭配 | 例句与翻译 | 语域 | | --- | --- | --- | --- | --- | | magnitude | /ˈmæɡnɪtjuːd/(≈麦格尼特尤德) n. | 量级、规模;the magnitude of, an order of magnitude | Few listeners grasp the magnitude of the shift. 很少有听众意识到这一转变的量级。 | 学术 / 通用 | | ballpark | /ˈbɔːlpɑːk/(≈宝尔帕克) adj./n. | 大致估计的;a ballpark figure, in the ballpark of | A ballpark figure would be 40,000 units a year. 大致估计是每年四万台。 | 商务 / 口语 | | sustained | /səˈsteɪnd/(≈瑟斯泰恩德) adj. | 持续的;sustained growth, a sustained fall | What matters is sustained improvement, not one good quarter. 重要的是持续改善,而不是一个季度的好看数字。 | 学术 / 商务 | | incremental | /ˌɪŋkrəˈmentl/(≈伊昂克若梅恩特尔) adj. | 渐进的、增量的;incremental gains, incremental change | The gains have been incremental: 1% to 2% a year. 改善是渐进式的:每年百分之一到二。 | 学术 / 商务 | | trajectory | /trəˈdʒektəri/(≈特若哲克特里) n. | 轨迹、走势;on an upward trajectory, change course | The trajectory matters more than the current level. 走势比当前水平更重要。 | 学术 / 媒体 | | saturation | /ˌsætʃəˈreɪʃn/(≈赛切雷希恩) n. | 饱和;reach saturation, market saturation | Smartphone sales have reached saturation in most rich markets. 智能手机销售在多数富裕市场已饱和。 | 商务 | | caveat | /ˈkæviæt/(≈凯维艾特) n. | 限定条件、保留;a caveat, with the caveat that | One caveat: the survey excludes informal workers. 一点保留:该调查不含非正式就业者。 | 学术 / 通用 | | displace | /dɪsˈpleɪs/(≈迪斯普莱斯) v. | 取代、挤走;displace workers, displace demand | Automation displaced routine tasks, not whole occupations. 自动化取代的是常规任务,不是整个职业。 | 学术 / 媒体 | | offset | /ˌɒfˈset/(≈奥夫塞特·短) v. | 抵消;offset the decline, be offset by | Falling prices were offset by higher volumes. 价格下跌被销量上升抵消。 | 商务 / 学术 | | broadly | /ˈbrɔːdli/(≈布劳德利·长) adv. | 大体上;broadly in line with, broadly unchanged | The findings are broadly in line with earlier estimates. 结论与早前估计大体一致。 | 学术 | 口语里最容易被忽视的是 broadly 与 roughly:它们不是填充词,而是给数字标注可信区间。删掉它们,你会把估计说成事实;用得太密,又会让整段听起来没底。 ## 三、语言点:口语数据的锚定结构与近似语分层 一段好的数据陈述有固定骨架:结论在前,数字在中,限定在后。数字本身不构成信息,数字加锚点才构成信息——"增长了 30%"几乎无意义,"三年里增长 30%,其中两年发生在补贴期内"才可被检验。近似语则是这套骨架的缓冲层,它告诉听众这个数是什么精度。 **规则** 1. 结论先行:先说这个数字要证明什么(Costs have stopped falling),再给数字。听众有了框架,数字才有位置。 2. 每个数字带三重锚:与什么比、哪一年、哪一群人。三者至少给两个。 3. 近似语按精度分层,不可越级混用: | 近似语 | 精度 | 适用 | 不适用 | | --- | --- | --- | --- | | about / around | ±5% 以内 | 已测得的数、口头引用 | 官方统计的精确值 | | roughly | 数量级内 | 心算、快速换算 | 需要担责的正式数字 | | in the order of | 十的幂级 | 大额、量级判断 | 小幅变化 | | north of / south of | 大于/小于某阈值 | 商务口语、趋势判断 | 学术写作 | | a ballpark figure | 明确承认是估算 | 预算、规划前期 | 已审计的数据 | 4. 一个数字只配一个近似语。~~approximately around 3.7%~~ 是双重模糊,听众会怀疑你根本不知道。 5. 限定语放在最后一句,且只放一个:The series stops in 2023, so the last two years are estimates. 两个以上限定会把结论稀释掉。 还有一个常被忽略的细节:数字的听辨高度依赖预期。演讲者说“up 30%, then 12%, then 4%”时,你会自动把三者当成同一指标的不同年份,因为结构给了你预期;反过来,如果他先报 30%,再报 41 天,中间不交代单位变化,你就会把 41 当成百分比。这也是为什么复述时必须把锚点说出来:锚点不只是给听众的,也是给你自己的,它逼你在开口前确认这个数字到底属于哪个序列。 | 弱(报数式) | 强(有锚有义) | 差别在哪 | | --- | --- | --- | | Sales grew 30%, then 12%, then 4%. | Growth has slowed for three years: 30%, then 12%, then 4%. The pattern, not any single year, is the story. | 前者是流水账,后者先给判断再给证据。 | | Unemployment is about 7.2%. | Unemployment is 7.2%, roughly double the rate before the reform. | 精确值加锚点,比模糊化更有说服力。 | | That's a huge number. | That is in the order of 2% of GDP—large enough to matter, small enough to reverse. | 用可比的量级替代形容词。 | **常见误用** 1. 连续报数不给停顿:三个数字连读,听众只记住最后一个。每个数字后停半秒,或加一句 which means。 2. 用 literally、basically 代替限定:~~it basically doubled~~。basically 是态度词,不是精度词。 3. 把百分比读成绝对数:a 3% rise 与 a rise of three percentage points 在口语里极易混淆,必须完整读出 percentage points。 4. 结尾不给 caveat:没有限定的数据陈述听起来像推销。一句 the sample is small 就能把语气拉回中立。 ## 四、句型与范例 | 句式 | 功能 | 例句与翻译 | | --- | --- | --- | | The headline number is …, but the story is in … | 先给总量,再指向结构 | The headline number is 4%, but the story is in who left the labour force. 总体数字是 4%,但关键在于谁退出了劳动力市场。 | | That is roughly double the rate of … | 用比较替代形容词 | That is roughly double the rate recorded before 2020. 这大约是 2020 年前水平的两倍。 | | What matters is not X but Y | 转移注意力到结构 | What matters is not the level but the trajectory. 重要的不是水平,而是走势。 | | In the order of N, which is … | 量级 + 含义 | The cost is in the order of 2% of GDP, which is roughly what the subsidy used to cost. 成本约为 GDP 的 2%,大致相当于过去补贴的规模。 | | Three numbers make the case: …, … and … | 明确预告结构 | Three numbers make the case: 30, 12 and 4. 三个数字说明问题:30、12 和 4。 | | Even if you discount X, the conclusion holds | 预先处理质疑 | Even if you discount the seasonal effect, the fall is sustained. 即使扣掉季节性因素,下滑仍是持续的。 | | One caveat before I go on: … | 主动交代局限 | One caveat before I go on: the series excludes the self-employed. 先说明一点:这组数据不含自雇者。 | | Put differently, … | 换一种说法加固 | Put differently, the market has stopped growing, not shrunk. 换句话说,市场停止增长,但没有萎缩。 | | I would not read this as … | 主动划定误读边界 | I would not read this as a recovery; it is a slower decline. 我不会把这读成复苏,它只是下滑变慢。 | | The figure that surprised me was … | 口语中引出重点 | The figure that surprised me was the 18% drop in repeat orders. 让我意外的是复购订单下降 18%。 | 这十句里,前三句决定听众是否跟得上,后两句决定他是否信你。中间的数字句可以按内容替换,但“结论—证据—限定”的次序不要动;一旦把限定提到开头,整段会变成防御姿态,听众会先怀疑你的立场,再听你的数据。 复述时不必十句全用。一句结论 + 三句数字 + 一句 caveat,2 分钟刚好,多说反而散。 ## 五、练习 **A 词汇与搭配(7 题)** 1. 改错:Sales grew approximately around 12%. 2. 填空(sustained / incremental):The improvement is real but ______: half a point a year. 3. 搭配:reach ______(饱和);offset the ______(下滑);on an upward ______(轨迹)。 4. 选词(magnitude / ballpark):Before the audit we only had a ______ figure; now we know the ______ of the shortfall. 5. 语域判断:哪句不能进学术段落?a) in the order of 3% b) north of 3% c) approximately 3% 6. 改错:Unemployment fell 2%. 7. 用 broadly 改写:The results are more or less the same as last year's. **B 句法、改写与压缩(6 题)** 1. 把流水账改成结论先行:Sales grew 30% in 2022, 12% in 2023 and 4% in 2024. 2. 给数字补锚:Costs fell 18%. 3. 把 "That's a huge number" 改成量级表述(自拟数字)。 4. 压缩到 30 词内,保留结论与一个数字:What we found, and this is after adjusting for seasonality and for the change in the sample, is that orders have fallen by about 9%, which is the third consecutive quarterly decline. 5. 加一句 caveat,使断言变稳:Remote work has permanently reduced office demand. 6. 把书面句改成口语可说句:A 12% reduction in headcount was recorded in the period under review. **C 输出(2 题)** 1. 五数字清单(表 5 行 × 5 列)+ 2 分钟复述录音。录音里必须出现:一句结论、三个数字、一句 caveat;不许出现 literally、basically。 2. 把复述转写为 120 词文字稿,标出每句的功能(结论/证据/限定/回指)。 ## 六、参考答案 **A** 1. 删一个:approximately 12% 或 around 12%。两个近似语叠加,等于告诉听众你连精度都没想清楚。 2. incremental。sustained 指"不间断",incremental 指"每次一点点",后半句的 half a point a year 指向后者。 3. saturation;decline;trajectory。三个都是固定搭配,换词就不地道。 4. ballpark;magnitude。前者是估算值,后者是规模本身。 5. b。north of 是商务口语,学术写作里应写 above 或 more than。 6. Unemployment fell 2 percentage points(或 fell by 2%)。原句无法判断是相对变化还是绝对变化,口语里必须补全。 7. The results are broadly in line with last year's. broadly 比 more or less 更书面,也更明确地表示"允许小幅偏差"。 **B** 1. Growth has slowed for three consecutive years: 30%, 12%, then 4%. 先给判断,数字退到冒号后当证据。 2. Costs fell 18% from their 2022 peak, and are now roughly level with 2019. 补上起点与对照年,数字才可检验。 3. That is in the order of 40 billion dollars—about 3% of GDP. 量级加占比,比 huge 有用得多。 4. After adjusting for seasonality, orders have fallen about 9%—the third consecutive quarterly decline. 保留调整说明、幅度、判断,删掉所有铺垫。 5. Remote work has reduced office demand, though the effect is concentrated in older buildings. 用 though 把断言收窄到可辩护的范围。 6. Headcount came down by around 12% over the period. 把被动名词化句还原成口语的主谓结构。 **C 评分维度** | 维度 | 合格 | 优秀 | | --- | --- | --- | | 结构 | 结论—数字—限定齐全 | 数字之间有递进,不是并列 | | 锚点 | 5 个数字中 ≥3 个带锚 | 5 个全带锚,且锚点不重复 | | 近似语 | 无双重模糊、无越级 | 精度与来源匹配,能解释为何这样模糊 | | 口语 | 2 分钟内无长停顿 | 有重音与停顿设计,关键数字前后各停半秒 | 参考转写(120 词):The short version is that hiring has cooled, not collapsed. Three numbers make the case. Postings are down 22% from the 2022 peak, which is roughly where they stood in 2019. Time-to-hire has stretched to 41 days, about a week longer than the pre-pandemic norm. And the share of roles filled internally has risen to 38%, in the order of ten points above 2019. Put differently, firms are replacing fewer people, and replacing them more slowly. One caveat: this series covers large employers only, so the gig economy is invisible here. ## 七、今日输出与自查 | 自查项 | 合格线 | 我的结果 | | --- | --- | --- | | 清单质量 | 5 个数字均支撑论点,装饰性数字已删 | | | 锚点 | 每个数字至少两个锚 | | | 近似语 | 全段无双重模糊,无学术/口语越级 | | | 复述结构 | 结论—三数字—限定齐全,2 分钟 ±15 秒 | | | 回听记录 | 3 条具体问题,指向结构与发音,不写"不流利" | | 产出文件:`第07周/Day04_五数字清单.md`(含清单表与 120 词转写)与 `Day04_复述录音.m4a`。录音文件名要能看出是第几周第几天,方便月底回听对比。 ## 八、明日预告 Day 5 是本周的写作主战场:300 词 synthesis,共识、分歧、空白三段,按主题组织而不是按来源。今天清单里的五个数字会成为那篇 synthesis 的证据;请把它们按"支持共识 / 暴露分歧 / 显示空白"预先分好类,明天落笔会快很多。注意:三段的主题句都不许出现作者名。