约9分钟阅读
你在写作中提出的每一个论点都需要支撑。你不能只抛出一个主张,就指望读者接受它。读者想知道:我凭什么相信这个?论据(evidence)就是你对这个问题的回答。它是你用来说服读者的材料,让读者相信你的主张成立、你的分析合理、你的结论站得住脚。
但并非所有论据都一样好,也并非所有论据都适用于每一种情况。不同类型的论据有不同的用途,高明的作者知道何时该用哪一种。本节介绍论据的主要类别,并说明每一类在什么时候最有效。
统计类论据用数字、测量结果、百分比和定量研究发现来支撑主张。在学术写作和职场写作中,它常被视为最有说服力的论据类型,因为它提供的是具体、可衡量的支撑。
统计数据给人以客观的感觉。它们提出精确、可衡量、可核实的主张。当读者看到来自可靠来源的具体数字时,论证就更有分量。数据还便于比较,而比较在议论文和分析性文章中至关重要。
统计数据如果脱离了语境,就可能误导人。"crime increased by 50%"(犯罪率上升了 50%)听起来很吓人,但如果实际数字是从 2 起变成 3 起,这种增长在统计上并不显著。作者必须如实地呈现统计数据,并提供足够的背景,让读者能准确地解读它们。
统计数据也无法解释某件事为什么会发生。它们显示的是相关性和规律,而不是因果关系。一项显示吃早餐的学生学业表现更好的统计数据,并不能证明是早餐带来了更好的表现。可能还有其他因素同时解释了这两种行为。
当你需要确立一个问题的范围或严重程度、展示随时间变化的趋势、比较不同群体或条件,或者证明可衡量的结果时,就使用统计数据。统计数据在政策论证、科学写作和商业语境中尤其有效。
专家证词(expert testimony)借助的是某个领域公认权威的知识和可信度。当一位顶尖研究者、一个专业组织或一家知名机构支持你的主张时,他们的声誉会为你的论证增添分量。
专家证词具有权威性。当一个主张有具备专业知识和职业资格的人背书时,读者更容易接受它。专家来源还表明,你的论证建立在已有的研究之上,而不是个人观点。
专家之间也会有分歧。在很多领域,严肃的学者持有相反的观点。如果同样可信的专家意见相左,引用其中一位并不能平息争论。作者应该承认专家意见存在分歧,并解释自己为什么认为其中一方更有说服力。
此外,光有权威并不能让论证成立。一位专家在自己的专业领域之外发表意见(比如一位物理学家评论经济政策),其分量不如在本领域之内发言的专家。
当你需要在技术性或专业性的话题上建立可信度,当读者可能对你的主张持怀疑态度,或者当权威人士的原话比你自己的转述更有说服力时,就使用专家证词。
例子是说明一般性主张的具体事例。它们向读者展示一个能证明你观点的具体案例、事件或情形,让抽象的观点变得可感可触。
例子让论证生动、令人难忘。统计数据给你数字,例子给你故事。一个选得好的例子能帮助读者看到你的论证在现实中是怎样成立的,这往往比单纯的抽象推理更有说服力。
一个例子并不能证明一条普遍规律。一家公司用某种策略取得了成功,并不意味着所有公司都能成功。例子是用来说明的,而不是用来下定论的。它们应该与其他形式的论据一起使用,而不是作为主张唯一的支撑。
例子也可能是精挑细选的。作者可能挑一个支持自己论证的例子,却忽略其他与之矛盾的例子。负责任的作者会承认自己的例子是个例外,或者承认存在反例。
当你需要把一个抽象的概念具体化,当读者可能不熟悉你的话题,或者当你想展示自己的论证在现实情境中如何成立时,就使用例子。在你已经给出统计数据或专家论据之后再举例子,效果尤其好,因为例子能让数字变得鲜活。
轶事(anecdote)是用来说明某个观点的简短故事或个人经历。与正式的例子不同,轶事往往涉及具体的个人及其经历,让论证带上人情味。
轶事能建立情感上的联系。它们为抽象的问题赋予一张人的面孔,帮助读者对受这个话题影响的人产生共情。在说服性的写作中,这种情感上的投入可能和逻辑推理同样重要。
从逻辑的角度看,轶事是最弱的一种论据。一个人的经历并不能证明一个普遍的趋势。"the plural of anecdote is not data"(轶事再多也不等于数据)这句话很好地概括了这一局限。轶事应该用来补充更有力的论据,而不是取代它们。
轶事也可能带有操纵性。动人的故事会分散读者的注意力,让他们忽略论证中的逻辑漏洞。有道德的作者用轶事来说明那些同样有数据和专家分析支撑的观点,而不是用它来绕过批判性思维。
用轶事来为文章写一个引人入胜的开头(引子),为数据密集的论证增添人情味,或者说明某个问题在现实中的影响。它们适合用在引言、结论中,也适合作为主体段内部的过渡。在学术写作中,要少用轶事,并且始终把它们与更有力的论据搭配使用。
类比(analogy)是把陌生的事物与读者已经了解的事物相比较,以此来解释它。严格来说,类比算不上论据,但它们是让复杂论证变得易懂、有说服力的有力工具。
类比让陌生的事物变得熟悉。它们把复杂或专业的概念与日常经验联系起来,帮助读者理解。一个精心打造的类比,能让论证在读者脑中「一下子想通」,这是单靠数据做不到的。
类比本质上是不完美的。没有两样事物是完全相同的,一个执意挑刺的批评者总能找到比较站不住脚的地方。类比可以澄清、可以说服,但不能证明。它应该用来补充论据,而不是替代论据。
类比也可能过度简化。把一个复杂的问题比作一个简单的问题,也许能让论证更易懂,但也可能歪曲真实的情况。
当你需要向非专业读者解释一个复杂的概念,当你想换一个角度重新审视一个问题,或者当你需要一种令人难忘的方式来概括你的论证时,就使用类比。类比在引言和结论中尤其有效。
最有力的论证会让多种类型的论据协同作用。一个段落可以先用一个统计数据开头,再用专家证词加以支撑,然后用一个具体的例子来说明:
Student mental health has deteriorated significantly since the widespread adoption of social media. A 2023 study published in the Journal of Youth Development found that teenagers who spend more than three hours daily on social media are twice as likely to report symptoms of anxiety and depression. Dr. Rachel Emerson, a psychology professor at Westfield University, attributes this correlation to "the constant social comparison and fear of missing out that these platforms engineer into their design." The experience of 16-year-old students at a suburban high school in Ohio illustrates this dynamic: after the school implemented a phone-free policy during class hours, counselor referrals for anxiety dropped by 30% within one semester.
这个段落结合了统计类论据(那项研究)、专家证词(Twenge 博士)和一个具体的例子(俄亥俄州的那所学校)。每一种论据都在强化其他论据,由此形成的论证比任何单一类型的论据都更有说服力。
高效的作者不会只依赖一种论据。他们会评估论证需要什么,并为具体的读者和目的选择最有说服力的论据类型。统计数据确立规模,专家证词确立可信度,例子让论证具体,轶事建立情感联系,类比让复杂的内容变得易懂。最好的论证会把这些类型交织在一起。
约8分钟阅读
了解有哪些类型的论据可用,只是第一步。更难的本领是选择:为你具体的论证、读者和语境挑选合适的论据。并非所有论据都同样有力,使用软弱的论据反而会损害你的可信度。本节教你用四条基本标准来评估论据:相关性(relevance)、可信度(credibility)、充分性(sufficiency)和时效性(recency)。
最常见的论据错误,是写进了有趣却无关的信息。只有当论据直接支撑你在这个段落中提出的具体主张时,它才是相关的。
选好一条论据后,问问自己:「如果我把这条论据删掉,我的主张会变弱吗?」如果答案是否定的,这条论据就与这个论证无关。它也许有趣,也许属实,但它不属于这里。
主张: Remote work increases employee productivity.
相关的论据:
无关的论据:
这六条论据都与远程办公有关,但只有前三条直接支撑关于生产效率的这个具体主张。其余几条也许能在别的段落里支撑别的主张,但它们不属于一个讨论生产效率的段落。
把论据放进段落之前,先检查它是否与段落的主题句相匹配。每一条论据都应该回答主题句提出的问题:
| 主题句 | 论据应该回答 |
|---|---|
| Exercise reduces anxiety. | 运动是怎样减轻焦虑的?减轻了多少? |
| Social media harms adolescent self-esteem. | 社交媒体是怎样损害自尊心的?通过什么机制? |
| Public libraries serve as community centers. | 图书馆提供了哪些社区功能? |
如果你的论据回答的是另一个问题,它就属于另一个段落。
可信的论据来自值得信赖的权威来源。论据的可信度直接影响论证的可信度。引用一份同行评审的期刊,比引用一篇匿名的博客文章更有分量。
用下面这些标准来判断一个来源是否可信:
作者的专业背景: 作者有资格谈论这个话题吗?他们有相关的教育背景、职业经验或研究资历吗?一位气候科学家谈气候变化,比一位政客或名人更可信。
出版方的声誉: 这个来源是由有声誉的机构发表的吗?同行评审的学术期刊、老牌新闻机构和公认的研究机构都有编辑标准,有助于保证准确性。自行发布的博客、社交媒体帖子和匿名网站则没有这样的标准。
研究方法的透明度: 对于研究类来源,它是否说明了数据是怎样收集和分析的?这项研究能被重复吗?交代了研究方法的研究,能让读者评估其结论是否成立。只给出结论、不解释结论从何而来的研究,应该持怀疑态度对待。
潜在的偏见: 这个来源在结论上是否有经济、政治或意识形态方面的利益?一项由饮料公司资助、研究糖对健康影响的研究,比一项由独立研究基金资助的研究更值得审视。偏见并不会自动让论据失效,但应该被承认并加以权衡。
同行评审与核实: 这条论据经过该领域其他专家的审查了吗?同行评审(peer review)的研究在发表之前已经由独立的学者评估过。这个过程并不完美,但它提供了一定程度的质量控制,而这是未经同行评审的来源所缺少的。
并非所有来源都有同等的分量。在学术写作和职场写作中,来源通常按可信度分为以下等级:
| 等级 | 来源类型 | 示例 |
|---|---|---|
| 最高 | 同行评审的研究、政府数据 | Journal of the American Medical Association、人口普查数据 |
| 高 | 知名机构、专家报告 | 世界卫生组织的报告、大学出版社的图书 |
| 中等 | 高质量的新闻报道、专业出版物 | The New York Times、Harvard Business Review |
| 较低 | 行业报告、倡导性组织 | 公司白皮书、非政府组织的出版物(可能有偏见) |
| 最低 | 个人博客、社交媒体、匿名来源 | 博客文章、推文、论坛评论 |
这个等级只是一个参考,而不是死板的规则。一篇调查充分的调查性新闻报道,可能比一项设计拙劣的学术研究提供更有力的论据。要看具体情况。
对具有以下特点的来源要保持警惕:
单独一条论据很少足以支撑一个主张。你需要多少论据,取决于你的主张有多大胆,以及你的读者有多怀疑。
主张越大胆,需要的论据就越多。这条原则有时被表述为 "extraordinary claims require extraordinary evidence"(非同寻常的主张需要非同寻常的证据)。
温和的主张(需要的论据较少): Regular exercise has been shown to improve mood.
这是被广泛接受的观点。一两项支撑研究就足够了,因为这个主张与常识和已有的研究一致。
大胆的主张(需要的论据较多): Thirty minutes of daily exercise is as effective as antidepressant medication for treating mild to moderate depression.
这挑战了传统的看法。它需要多项同行评审的研究(最好是一项元分析)、专家证词,并仔细说明这个主张在哪些具体条件下成立。
没有放之四海而皆准的公式,但下面这些准则会有帮助:
对于单个主体段: 两到三条论据通常就足够了。只有一条论据,主张会显得单薄;超过四条,会让读者应接不暇,也没有留下分析的空间。
对于整篇文章的论证: 每个主体段都应该有自己的论据,而且只要可能,不同的段落应该引用不同的来源。一篇从头到尾只引用一个来源的文章——无论引用了多少次——都会显得研究不足。
对于有争议的主张: 提供来自多个独立来源的论据。如果三项互不相关的研究得出了相同的结论,论证就比只有一项研究支持时更有力。
给出论据之后,想象一位持怀疑态度的读者在问:「就这些吗?我还没被说服。」如果你能听到这样的反驳,就说明你需要更多的论据。继续补充支撑,直到一位讲道理的读者会接受这个主张为止。
支撑不足: 提出了一个很强的主张,却只给出一个例子或一个统计数据。这会让读者无法信服。
支撑过度: 堆砌了太多论据,让段落变成一堆没有分析的数据。没有解释的论据只是噪音。
依赖单一来源: 所有论据都完全依赖一个来源。如果这个来源有缺陷,你的整个论证就会崩塌。要让来源多样化。
在很多领域,论据是有保质期的。20 年前的数据可能无法反映当前的情况。使用过时的论据会削弱你的论证,在变化迅速的话题上尤其如此。
科技: 一项 2012 年关于社交媒体使用情况的研究,在今天基本上已经不相关了。平台、用户构成、算法和使用模式都已经发生了巨大的变化。
科学与医学: 随着新研究的出现,医学指南在不断更新。引用一项已被取代的研究,可能得出不准确甚至有害的结论。
经济与政治: 经济状况、政策和政治格局都会随时间而变化。一项疫情前研究中的就业数据,并不能反映疫情后的现实。
人口统计: 人口统计数据、移民模式和社会趋势在几十年间会发生显著变化。始终使用可获得的最新数据。
历史分析: 如果你写的是第一次世界大战的起因,1914 年的一手资料并不「过时」,反而是必不可少的。
已确立的科学原理: 热力学定律没有改变。基础的科学原理不需要最新的来源。
经典文学与哲学: 亚里士多德的《修辞学》并不因为写于公元前 4 世纪就过时了。在文学和哲学分析中,来源的年代往往无关紧要。
里程碑式的研究: 有些研究设计得非常出色、影响力极大,以至于几十年后依然具有权威性。米尔格拉姆服从实验(1961 年)至今仍在心理学中被引用,因为后来的研究都没有推翻它的核心发现。
在文章中使用任何一条论据之前,都用这四个问题检验一下:
只要其中任何一个问题的答案是否定的,就要么去找更好的论据,要么调整你的主张,让它与论据实际能支撑的程度相匹配。优秀的作者从不强迫软弱的论据去承担重任,而是为这项任务找到合适的论据。
约9分钟阅读
找到有力的论据固然重要,但这只完成了一半。另一半是把论据自然地融入你的文章。融合得不好,文章就会显得断断续续、支离破碎,或者像是别人观点的拼贴。融合得好,论据就会被编织进你自己的论证之中,让每一个来源都服务于你的声音,而不是取代它。
本节讲解引用来源材料的三种主要方法——直接引用、转述和概括——以及用来引出论据、并把它与你的论证联系起来的引导语。
直接引用(direct quote)是一字不差地照录来源的原话。你逐字复制原文,并加上引号。
直接引用要有选择地使用。只在原文措辞必不可少的情况下才用:
措辞本身很重要。 如果作者的表达特别有力、精确或令人难忘,直接引用能保留这种效果:
Martin Luther King Jr. argued that "injustice anywhere is a threat to justice everywhere," a principle that remains central to modern civil rights advocacy.
你在分析具体的措辞。 在文学分析或修辞分析中,原话很重要,因为你考察的正是作者怎样使用语言:
Orwell's choice of the word "vaporized" to describe the elimination of political dissidents reveals the regime's desire to erase not just the person but any evidence that they ever existed.
来源具有独特的权威性。 当一位顶尖专家、一份官方文件或一项法律裁决作出明确的表态时,直接引用它能让你的论证获得这种权威的分量:
The Supreme Court ruled that the policy "constitutes an unconstitutional infringement on the right to free expression," leaving no room for ambiguity about its legal standing.
在以下情况下,避免直接引用:
学生写作中的一个常见错误是过度引用:对本该转述的常规信息使用直接引用。如果一个来源说 "approximately 40% of food produced globally is wasted each year",你没有必要引用这句话。重要的是数据,而不是措辞。转述为:Nearly 40% of the world's food production goes to waste annually.
引语必须在语法上与你的句子契合。它读起来应该是你行文中自然的一部分,而不是生硬的插入。
生硬的融入: The study found that. "Remote workers were 13% more productive than their office counterparts." This shows productivity increases.
自然的融入: The study found that "remote workers were 13% more productive than their office counterparts," suggesting that the traditional office environment may actually hinder output.
自然融入的版本把引语嵌进了句子结构,并紧接着给出分析。
在学术写作中,超过大约 40 个词的引语通常采用独立引文(block quote)格式——从正文缩进,不加引号。不过,在大多数文章写作中(尤其是 IELTS、TOEFL、AP 这类限时考试),应该完全避免长引语。它们占用字数,却不能展示你的分析能力。如果你需要提到一段长文,就转述或概括它,只引用其中最重要的短语。
转述(paraphrasing)是用你自己的措辞和句子结构重新表述别人的观点。与引用不同,转述不使用原文的措辞;与概括不同,转述涵盖的细节程度与原文相同。
转述是融入论据最重要的技能,因为它能让你在引用来源材料的同时,保持自己的声音,并展示你对内容的理解。有效的转述向读者表明,你不只是找到了相关的信息,而且消化并理解了它。
第 1 步:阅读并理解原文。 不要一边看着原文一边转述。读完这段文字,合上它或把目光移开,然后用你自己的话写出这个意思。
第 2 步:同时改变措辞和句子结构。 只是把几个词换成近义词并不是转述。这叫拼凑式改写(patchwriting),是一种学术不端行为。你必须从根本上重构这个句子。
原文: "The proliferation of misinformation on social media platforms has eroded public trust in traditional media institutions."
拼凑式改写(不可接受): The spread of misinformation on social media sites has damaged public trust in traditional media organizations.
这个版本把 "proliferation" 换成了 "spread",把 "platforms" 换成了 "sites",把 "eroded" 换成了 "damaged",但句子结构完全一样。它与原文太接近了。
真正的转述(可接受): Traditional media organizations are losing credibility with the public, in part because social media has made it easy for false information to spread widely and rapidly, blurring the line between reliable journalism and unverified claims.
这个版本彻底重构了这个意思:因果关系保留了下来,但句子结构、用词和重点都是全新的。
第 3 步:把转述与原文对照检查。 写完转述后,把它与原文比较一下。如果有任何连续三个或以上的词与原文相同,就进一步修改。
第 4 步:注明来源。 转述仍然需要注明出处。即使你用的是自己的话,这个观点最初也是别人提出的。
照搬句子结构: 保留相同的语法结构、只替换词语,不是转述。
歪曲原意: 你的转述必须准确地表达来源的观点。不要为了更契合你的论证而夸大、弱化或改变作者的观点。
忘记注明出处: 因为转述用的是你自己的话,所以很容易忘记这个观点仍然属于别人。一定要为转述的内容注明来源。
概括(summarizing)是把一段较长的文字、一篇文章或一项研究浓缩成它的要点。与涵盖同等细节程度的转述不同,概括只抓住主要观点。
在以下情况下使用概括:
只关注主要观点。 概括应该抓住一个来源「说了什么」以及「为什么重要」,而不是每一个支撑细节。
原文(150 词): A longitudinal study conducted at three urban hospitals tracked 500 patients with Type 2 diabetes over five years. Participants were randomly assigned to either a standard care group or an intervention group that received weekly nutritional counseling and a personalized exercise plan. After five years, the intervention group showed a 35% reduction in HbA1c levels, a 28% reduction in cardiovascular events, and a 40% reduction in diabetes-related hospitalizations. The study concluded that sustained lifestyle intervention is more effective than medication alone in managing Type 2 diabetes and recommended that healthcare systems integrate nutritional counseling and exercise programs into standard diabetes care protocols.
概括(40 词): A five-year study of 500 diabetes patients found that those who received nutritional counseling and personalized exercise plans had significantly better health outcomes than those receiving standard care, suggesting that lifestyle intervention should be integrated into diabetes treatment.
这个概括保留了核心的发现和建议,同时略去了具体的百分比和研究方法上的细节。
在学术写作中,你常常需要概括一批研究,而不是单个来源:
Multiple studies over the past decade have consistently found that bilingual individuals outperform monolinguals on tasks requiring cognitive flexibility, attention switching, and inhibitory control, leading researchers to describe bilingualism as a form of cognitive exercise that strengthens executive function.
这句话概括了整个研究领域,而没有引用任何一项具体的研究。当你想在呈现具体论据之前先确立一种普遍共识时,这样做是恰当的。
引导语(signal phrase)用来引出来源材料:它指明来源,并表明你是怎样使用它的。引导语是你自己的论证与别人的论据之间的桥梁。
没有引导语,论据就会毫无背景地出现在你的文章里。读者不知道是谁说的、为什么重要,也不知道它与你的论证有什么联系。引导语通过以下方式解决这个问题:
引导语通常包括作者的名字(或机构的名称),以及一个描述来源做了什么的动词:
[Author/Source] [verb] that [evidence].
例子:
你在引导语中选择的动词,传达了你与这个来源的关系。不同的动词带有不同的含义:
| 动词 | 含义 |
|---|---|
| argues, claims, contends, asserts | 来源在提出一个可争论的主张 |
| demonstrates, shows, proves, establishes | 来源提供了令人信服的论据 |
| suggests, indicates, implies | 论据指向某个方向,但并不确定 |
| notes, observes, points out, states | 中立地呈现信息 |
| acknowledges, concedes, admits | 来源承认了一个可能削弱其立场的观点 |
| warns, cautions | 来源在发出警示 |
| disputes, challenges, rejects, refutes | 来源反对另一种立场 |
在 "Martinez argues" 和 "Martinez demonstrates" 之间作出选择,表明的是你认为这条论据仅仅是有意思,还是真正令人信服。这种细微差别很重要。
避免反复使用同一个引导语。"According to..." 很有用,但如果每一段都出现,就会变得单调。变换一下你的写法:
搭配直接引用: Dr. Elena Vasquez, a leading epidemiologist, warns that "the next pandemic is not a question of if but when, and our current surveillance systems are dangerously underprepared."
搭配转述: Dr. Elena Vasquez, a leading epidemiologist, warns that global pandemic surveillance systems remain inadequate despite the lessons of recent outbreaks.
搭配概括: In her comprehensive review of global health preparedness, Dr. Elena Vasquez concludes that current surveillance systems are insufficient and recommends a coordinated international monitoring network.
在每个例子中,引导语都做着同样的工作:指明来源,确立可信度(leading epidemiologist),并用一个动词(warns、concludes)表明来源所作贡献的性质。
融入论据最可靠的框架之一,是「三明治法」(the sandwich method)。它把每一条论据分成三层来安排:
例:
The economic benefits of immigration extend beyond job creation to fiscal contributions. [上层面包——引出主张] A comprehensive study by the National Academy of Sciences found that first-generation immigrants and their descendants contribute an average of $259,000 more in taxes over their lifetimes than they receive in government benefits. [夹心——论据] This net positive fiscal impact directly contradicts the common misconception that immigration is an economic burden, and it suggests that restrictive immigration policies may actually reduce government revenue. [下层面包——分析]
三明治法确保论据永远不会孤立存在。它总是被引出、总是有背景、总是得到分析。
约9分钟阅读
光呈现论据是不够的。学生写作和职场写作中最常见的错误,是把论据当成不言自明的东西——把一个统计数据、一句引语或一个例子往段落里一放就继续往下写,好像论据会自己说话。它不会。没有分析的论据,就像没有律师的证人:事实摆在那里,却没有人向陪审团解释它们意味着什么。
本节专注于分析论据这项关键技能——也就是把一堆事实变成一个有说服力的论证的那个「so what?」因素。
每给出一条论据,都想象读者在问:「So what?(那又怎样?)这为什么重要?它证明了什么?」你的分析就是对这个问题的回答。
看看一位作者给出了论据却没有加以解释时会发生什么:
School uniforms should be required in public schools. A study by the University of Houston found that after implementing a mandatory uniform policy, attendance rates increased by 5% and disciplinary referrals decreased by 12%.
这条论据既相关又可信,但作者没有把它与主张联系起来。读者只能自己去琢磨其中的联系。校服为什么能提高出勤率?违纪转介的减少怎样支撑「应该强制穿校服」这一主张?这条论据对于更大的论证究竟证明了什么?
再看同样的论据加上分析之后:
School uniforms should be required in public schools because they reduce distractions and create a more focused learning environment. A study by the University of Houston found that after implementing a mandatory uniform policy, attendance rates increased by 5% and disciplinary referrals decreased by 12%. These improvements suggest that when students are not preoccupied with clothing choices — and the social pressures, competition, and bullying that accompany them — they are more likely to attend school regularly and engage in fewer conflicts. The reduction in disciplinary referrals is particularly significant because it indicates that uniforms do not merely suppress visible misbehavior but actually reduce the social tensions that cause disruptions in the first place.
分析做到了三件单凭论据做不到的事:
分析论据的核心本领,是在论据所显示的内容和你所论证的内容之间建立明确的联系。作为作者,你可能觉得这些联系显而易见,但对读者来说,它们从来都不是显而易见的。你必须把它们说清楚。
一个把论据与主张联系起来的实用公式:
论据 → 解读 → 与主张的联系
下面看看它在实践中是怎样运作的:
主张: Access to green spaces improves mental health in urban populations.
论据: A longitudinal study in Barcelona tracked 3,500 residents and found that those living within 300 meters of a park reported 25% lower rates of depression and anxiety compared to those living more than one kilometer from green spaces.
解读: The proximity effect suggests that green spaces provide ongoing, cumulative mental health benefits rather than one-time relief. Residents near parks are more likely to use them regularly for exercise, social interaction, and quiet reflection — all of which are established interventions for depression and anxiety.
与主张的联系: This finding supports the argument that urban planners should prioritize equitable green space distribution, particularly in low-income neighborhoods where both park access and mental health resources are typically scarce.
注意,分析是怎样从论据向外层层展开的:先解读数据意味着什么,再把这种解读与关于城市规划的更大论证联系起来。
研究下面这些改写,看看分析怎样把软弱的段落变得有力。
修改前(有论据,没分析):
Carbon taxes are effective at reducing emissions. Sweden introduced a carbon tax in 1991 at $26 per ton, which has since increased to over $130 per ton. Between 1990 and 2020, Sweden reduced its carbon emissions by 29% while its GDP grew by 85%.
论据很有力,但作者没有解释它为什么重要。读者只能自己得出结论。
修改后(有论据,有分析):
Carbon taxes are effective at reducing emissions without sacrificing economic growth. Sweden introduced a carbon tax in 1991 at $26 per ton, which has since increased to over $130 per ton. Between 1990 and 2020, Sweden reduced its carbon emissions by 29% while its GDP grew by 85%. This simultaneous reduction in emissions and growth in GDP directly refutes the common objection that carbon pricing will destroy the economy. Sweden's experience demonstrates that when carbon taxes are implemented gradually and revenues are reinvested in clean energy infrastructure, businesses adapt by finding more efficient production methods rather than simply passing costs to consumers. The Swedish model suggests that the perceived trade-off between environmental protection and economic prosperity is false — a lesson that should inform the debate in countries still resisting carbon pricing.
分析解释了瑞典的数字为什么重要(它们驳倒了一个常见的反对意见)、这项政策怎样起作用(逐步实施并把收入再投资),以及它对其他国家有什么意义(所谓的取舍并不存在)。
修改前(有论据,没分析):
Class size affects student learning. A study in Tennessee randomly assigned students to small classes (13-17 students) or regular classes (22-25 students) from kindergarten through third grade. Students in smaller classes scored higher on standardized tests in every grade level.
作者给出了论据就停下了。读者知道了结果,却不明白它的意义。
修改后(有论据,有分析):
Reducing class sizes is one of the most consistently effective interventions for improving student achievement, particularly in the early grades. A landmark study in Tennessee randomly assigned students to small classes (13-17 students) or regular classes (22-25 students) from kindergarten through third grade. Students in smaller classes scored higher on standardized tests in every grade level. The advantage of this study is its design: because students were randomly assigned, the results cannot be attributed to differences in family income, school quality, or teacher ability. The smaller classes allowed teachers to provide more individualized attention, identify struggling students earlier, and spend less time on classroom management. These effects were especially pronounced among students from low-income families, who gained the equivalent of three additional months of learning per year — a finding that suggests class size reduction could be a powerful tool for closing the achievement gap between wealthy and disadvantaged students.
分析涉及了研究方法(随机分配排除了混杂因素),解释了作用机制(更多个别关注、更少课堂管理时间),并得出了更大的意义(缩小成绩差距)。
修改前(有论据,没分析):
Processed food is linked to health problems. A study published in the British Medical Journal followed 100,000 adults for five years and found that a 10% increase in ultra-processed food consumption was associated with a 12% increase in cancer risk.
这是一个很有分量的统计数据,却在没有背景、没有解读的情况下被摆了出来。
修改后(有论据,有分析):
The health risks of ultra-processed food extend far beyond weight gain to include significantly elevated cancer risk. A study published in the British Medical Journal followed 100,000 adults for five years and found that a 10% increase in ultra-processed food consumption was associated with a 12% increase in cancer risk. The scale of this study — 100,000 participants over five years — gives its findings considerable weight, and the dose-response relationship (more processed food correlates with proportionally higher risk) strengthens the case for a causal connection rather than mere correlation. This evidence is particularly alarming given that ultra-processed foods now constitute more than 50% of calories consumed in many developed nations. If even a fraction of the observed cancer risk increase is causal, the public health implications are staggering, and they argue for regulatory interventions similar to those applied to tobacco: warning labels, advertising restrictions, and taxation.
分析评估了这项研究的分量(规模、剂量—反应关系),为研究发现提供了背景(加工食品的消费比例),并得出了政策上的意义(像监管烟草那样加以监管)。
并非所有论据都需要同样深度的分析。你提供多少分析,应该与论据的重要程度,以及它与主张之间联系的复杂程度相匹配。
当论据与主张之间的联系一目了然时,一两句分析就够了:
The city's investment in public transit reduced average commute times by 15 minutes, demonstrating that infrastructure spending directly improves residents' quality of life.
这种联系很清楚,几乎不需要解释。
当论据很复杂、出人意料,或者是你论证的核心时,就要提供更透彻的分析:
Although Finland spends less per student than the United States, Finnish students consistently outperform American students on international assessments. This counterintuitive finding suggests that education quality is not primarily a function of funding but of how resources are allocated. Finland invests heavily in teacher training and professional autonomy while spending relatively little on standardized testing and administrative overhead. The contrast with the American model — which spends billions on testing while underpaying teachers — reveals a fundamental misalignment between spending priorities and educational outcomes. Money matters, but it matters most when directed toward the people and practices that directly affect learning.
这种更深入的分析是恰当的,因为这条论据挑战了一个常见的假设,需要较长的解释才能与论证联系起来。
不同类型的论据,通常需要不同的分析角度:
| 论据类型 | 分析重点 |
|---|---|
| 统计数据 | 这些数字意味着什么?它们显著吗?它们暗示了什么? |
| 专家引语 | 这位专家的观点为什么重要?它怎样支撑你的论证? |
| 例子 | 这个案例说明了什么?它有代表性,还是个例外? |
| 轶事 | 这个故事说明了什么更大的观点?它与数据有什么联系? |
| 类比 | 这个比较在哪些方面成立?在哪里站不住脚? |
有些作者以为自己在分析,其实只是把论据换个说法:
论据: 75% of employees reported higher job satisfaction after switching to a four-day workweek.
复述(不是分析): This shows that three-quarters of workers were more satisfied with their jobs after working four days instead of five.
这什么也没有增加。它说的和论据已经说过的一模一样。
真正的分析: This overwhelming majority suggests that the benefits of a compressed schedule — more personal time, reduced burnout, greater autonomy — outweigh the potential drawbacks of longer daily hours. Notably, the finding that satisfaction increased even when total weekly hours remained the same indicates that the structure of the workweek, not just its length, significantly affects employee well-being.
得出的结论远远超出了论据所能支撑的范围:
论据: A study found that students who took handwritten notes performed better on conceptual questions than students who typed notes on laptops.
过度概括: This proves that technology has no place in the classroom and that we should return to traditional teaching methods.
这项研究讲的是做笔记,而不是所有课堂技术。这种分析从一个具体的发现,一下子跳到了一个笼统的主张。
恰当的分析: This finding suggests that the physical act of handwriting forces students to process and condense information rather than transcribing it verbatim, which may promote deeper understanding. However, the advantage applies specifically to conceptual learning and may not extend to other uses of technology in the classroom.
只呈现支持你主张的论据,而忽视那些使之复杂化或与之矛盾的论据:
优秀的作者会在分析中承认局限和反面论据。这不会削弱你的论证;相反,它能展示你在学术上的诚实,向读者表明你已经考虑了全局,从而加强你的论证。
While the Barcelona study strongly supports the mental health benefits of urban green spaces, it is worth noting that access to parks alone does not address the structural causes of mental health disparities in low-income communities, such as housing instability, food insecurity, and inadequate healthcare. Green spaces should be viewed as one component of a comprehensive approach, not a standalone solution.
有些作者对论据的分析很有意思,却忘了把分析与最初的主张联系起来:
分析的结尾,一定要回到段落的主题句或文章的论点句上。永远不要让读者去猜论据和分析与你的主要论证有什么关系。
约7分钟阅读
本节提供在文章中融入和分析论据的速查工具。把这些模式当作模板,用常见错误清单来自我诊断,在写作和修改时用速查表随时参考。
下面这些段落层面的模式,展示了怎样在主体段中安排论据。每种模式都有不同的用途。选择最适合你的论证和论据的那一种。
这是最常见、用途最广的模式。它适用于任何类型的论据和任何文章形式。
结构:
模板:
[Claim]. [Context for the evidence]. [Evidence from source]. [Analysis — what does this evidence mean?]. [Connection — how does this support the thesis?].
例:
Early childhood education produces long-term economic benefits that far exceed its costs. The Perry Preschool Project, one of the longest-running educational studies, tracked 123 children from age 3 through age 40. Participants who received two years of high-quality preschool earned 14% more annually, were 20% more likely to own a home, and were 50% less likely to have been arrested five or more times compared to the control group. These outcomes translate into a return of $7 to $12 for every dollar invested in the program, according to economists who analyzed the data. When the costs of preschool are weighed against the savings in welfare, criminal justice, and increased tax revenue, early childhood education is not merely a social good but a sound financial investment.
当一个主张需要多个来源的支撑时,这种模式能让你高效地层层叠加论据。
结构:
模板:
[Claim]. [Evidence 1 and what it shows]. [Evidence 2 and what it adds]. [Evidence 3 and what it contributes]. [Synthesis — together, what do these sources prove?].
例:
The four-day workweek improves employee performance across multiple industries. A Microsoft Japan trial in 2019 found that productivity increased by 40% after implementing Friday closures, with employees completing the same volume of work in fewer hours. An Icelandic government study of 2,500 workers across offices and hospitals found that productivity remained the same or improved in 86% of workplaces that reduced hours. More recently, a 2022 pilot in the United Kingdom involving 61 companies reported that 92% chose to continue the four-day week permanently after the trial period. The consistency of these results across countries, industries, and organizational sizes suggests that the five-day workweek is not optimized for productivity but is rather a historical convention that persists out of inertia.
这种模式先承认相反的论据,再提出你自己的论据,以此展示学术上的诚实,并加强你的论证。
结构:
模板:
[Critics/opponents argue that...]. [However, evidence shows...]. [This is more convincing because...]. [Therefore, the claim stands].
例:
Critics of renewable energy often point to the intermittency problem: solar panels do not generate electricity at night, and wind turbines are idle on calm days. This is a legitimate concern, but recent advances in grid-scale battery storage have largely addressed it. Tesla's 150-megawatt battery installation in South Australia has demonstrated that batteries can respond to grid fluctuations faster than traditional power plants, providing backup power within milliseconds of a supply drop. Combined with the fact that the cost of lithium-ion batteries has fallen by 89% since 2010, the intermittency argument is rapidly becoming obsolete. While storage technology is not yet sufficient to replace all fossil fuel baseload power, the trajectory of cost reduction and capacity improvement makes full grid reliability achievable within the next decade.
有些论证最好用一个深入分析的详细例子来支撑,而不是用多条简短的论据。
结构:
模板:
[Claim]. [Introduction of specific case]. [Key details of the case]. [What this case demonstrates]. [How this connects to the broader argument].
例:
Investing in preventive public health measures saves significantly more money than treating diseases after they occur. The case of lead paint removal in Rochester, New York illustrates this principle clearly. In the early 2000s, Rochester had one of the highest rates of childhood lead poisoning in the United States, with an estimated 3,000 children affected annually. The city launched a comprehensive lead abatement program that spent $50 million over a decade removing lead paint from housing, testing children, and educating landlords. By 2015, childhood lead poisoning rates had dropped by 80%, and the city estimated savings of over $200 million in reduced special education costs, healthcare expenses, and criminal justice costs associated with lead exposure's neurological effects. Rochester's experience demonstrates that the upfront cost of prevention, while politically difficult to justify, is dwarfed by the long-term savings in treating preventable conditions.
这种模式用相互对照的论据来加强主张,展示有和没有所讨论的变量时分别会发生什么。
结构:
模板:
[Claim based on comparison]. [Case A shows...]. [In contrast, Case B shows...]. [This contrast demonstrates...]. [Connection to thesis].
例:
Countries that invest in universal pre-kindergarten programs see measurably better educational outcomes than those that do not. France, which has offered free preschool to all children aged 3-5 since the 1980s, has a 99% enrollment rate and consistently ranks above the OECD average in early literacy assessments. The United States, where pre-kindergarten access varies dramatically by state and income level, has a preschool enrollment rate of only 54% among 3-year-olds, and American children from low-income families enter kindergarten an average of 12 months behind their wealthier peers in language and cognitive development. This contrast suggests that universal access, rather than targeted programs for specific populations, is the more effective approach to early childhood education, because it prevents gaps from forming rather than attempting to close them after they have already widened.
用这份诊断清单来发现并改正文章中的论据问题。
表现: 一个统计数据、一句引语或一个例子被丢进段落里,却没有解释它意味着什么、为什么重要。
Social media affects mental health. A study found that teenagers who use Instagram more than two hours daily are 30% more likely to report depression symptoms. Another study found that cyberbullying victims are twice as likely to self-harm.
作者给出了两条有力的论据,却一条也没有分析。这个段落读起来像一张事实清单,而不是一个论证。
怎么改: 每一条论据之后,至少加一句分析,解释这条论据证明了什么,以及它怎样与你的主张相联系。
表现: 段落大部分由引语构成,作者自己写的内容寥寥无几。
According to Smith, "renewable energy is the future of power generation." Jones agrees, stating that "solar and wind power have become cost-competitive with fossil fuels." Furthermore, the IPCC notes that "a rapid transition to clean energy is essential to limiting warming to 1.5 degrees."
作者成了一个汇编者,而不是一个论证者。他们自己的声音消失了。
怎么改: 大部分论据用转述,把直接引用留给那些格外有力或具有权威性的原话。在各条论据之间加上你自己的分析。
表现: 论据很有意思,呈现得也很好,却并不真正支撑段落中提出的那个具体主张。
主张: Public libraries improve literacy rates. 给出的论据: Public libraries hosted over 50,000 community events last year, including author talks, film screenings, and crafting workshops.
这条论据讲的是社区活动,而不是读写能力。它也许能支撑另一个主张(图书馆是社区中心),但支撑不了这一个。
怎么改: 在使用论据之前,先确认它直接支撑你主题句中的那个具体主张。如果它支撑的是另一个主张,就把它移到合适的段落里。
表现: 作者用一个个人故事或单个案例,作为一个普遍性主张的主要支撑。
Standardized tests do not measure intelligence. My cousin scored poorly on the SAT but graduated top of her class in college and is now a successful engineer.
一个人的经历并不能证明一条普遍原则。这个轶事也许能说明这个主张,但单靠它支撑不了。
怎么改: 用轶事来说明那些同样有统计论据和专家分析支撑的观点。轶事增添人情味,数据提供证明。
表现: 作者引用几十年前的论据,来支撑一个关于当前情况的主张。
Social media is not harmful to teenagers. A 2008 study found that online communication helped adolescents maintain friendships and develop social skills.
2008 年的社交媒体(早期的 Facebook,没有 Instagram,也没有 TikTok)与今天的社交媒体有着本质的不同。这条论据太旧了,已经不相关。
怎么改: 对于涉及科技、医学、经济和社会趋势的话题,使用近五年内的论据。如果必须引用较早的研究,就承认它的年代,并解释它为什么依然相关。
| 情况 | 最佳做法 |
|---|---|
| 原文措辞格外有力 | 直接引用 |
| 你在分析具体的用词 | 直接引用 |
| 信息是事实性的、常规的 | 转述 |
| 你需要保持自己的声音 | 转述 |
| 你在提到一整项研究或一整篇文章 | 概括 |
| 你在提供背景信息 | 概括 |
| 主张很温和、被广泛接受 | 1–2 条论据 |
| 主张很大胆或有争议 | 3 条及以上、来自独立来源的论据 |
| 与主张的联系显而易见 | 浅层分析(1 句) |
| 联系很复杂或出人意料 | 深层分析(2–4 句) |
| 中立地呈现论据 | states, reports, notes, observes, finds, identifies |
|---|---|
| 呈现你认为令人信服的论据 | demonstrates, establishes, proves, confirms, shows |
| 呈现一个可争论的主张 | argues, claims, contends, asserts, maintains, proposes |
| 呈现尚不确定的发现 | suggests, indicates, implies, hints |
| 呈现反对意见 | disputes, challenges, rejects, counters, questions, criticizes |
| 表示赞同 | agrees, supports, confirms, corroborates, echoes |
| 表示让步 | acknowledges, concedes, admits, recognizes, grants |
对文章中的每一个主体段都使用这份清单:
在一个结构良好的论据段中:
如果论据多于分析,说明你分析不足;如果主张多于论据,说明你支撑不足。分析永远应该是占比最大的部分,因为你的论证就存在于分析之中。