IB TOK 知识问题 论证 反方 概念 展评 论文
TOK 的两项评估任务都从一句知识问题开始，也都以论证收尾。展评要三件物件对同一条 IA prompt，论文要在 1600 词里回应一道官方题。写得顺的学生往往在提问那一步就想清楚了要证明什么。写得吃力的学生常常把例子当装饰，写到一半才发现自己没有主张。下面把知识问题的判定、论证的骨架，以及两项任务各自的要求分开说。
Both TOK tasks begin with a knowledge question and end with an argument. Students who write smoothly usually know what they are proving before they collect examples, while the ones who struggle treat examples as decoration and discover halfway through that no claim was ever made. The sections below separate the anatomy of a usable knowledge question from the anatomy of a usable argument, then apply both to the exhibition and the essay.
知识问题与学科问题的分界
TOK 的整个课程围绕知识问题展开，IBO 的《Theory of knowledge guide》（2020 年 2 月发布，2022 年首次评估）把这类问题定义为关于知识本身的可争议问题。它问知识怎样产生、怎样获得、怎样传播、怎样被使用。它也问谁知道、谁不知道、谁有资格判定答案。学科问题问的是具体内容，比如某条定律的公式怎么写，某场战争发生在哪一年。知识问题退后一层，问我们凭什么相信这条定律，问这场战争的叙述为什么被反复改写。很多学生以为 TOK 就是换一种方式重讲学科内容，其实两者的提问层级完全不同。课堂上老师多追问一句「你怎么知道」，往往就是把学科问题推向知识问题的动作。判断办法很朴素：能靠翻教材索引找到答案的，多半是学科问题；几个受过训练的成年人能给出互相冲突又都站得住的答案，才落在 TOK 的地界上。这门课是 DP 的核心要素，IBO 要求学校至少投入 100 小时教学时间。
TOK is built around knowledge questions, which the IBO guide, published in February 2020 with first assessment in 2022, defines as contestable questions about knowledge itself. They ask how knowledge is produced, acquired, shared and used, and who is entitled to settle the answer. A subject question asks what a formula says or when a war began, whereas a knowledge question steps back and asks why the formula is trusted or why accounts of the war keep being rewritten. Many students assume TOK is their subject content retold; the two differ in the level at which they ask. A quick test works: if a textbook index answers it, it is probably subject content, and if trained adults can give conflicting answers that both hold up, it belongs to TOK. The course is a DP core element, and the IBO requires schools to devote at least 100 teaching hours to it.
判定一个知识问题是否站得住的四个条件
IBO 的考纲给出三个标记：可争议、关于知识本身、用 TOK 的概念与术语而不是学科术语。三个标记之外还有一条常被忽略的条件，问题必须开放，不能把正确答案预先藏在措辞里。写「科学方法是否比历史方法更可靠」，这题看着开放，实际预设了可靠性只有一种量法。改成「不同领域判定可靠性的标准能不能互相比较」，争议才真正打开。学生常把「是不是」「应不应该」当成开放，其实开放体现在答案可以沿着几条不同路径通向不同结论。题干里出现「最重要」「唯一」这类词，答案往往已经写进问题里了。稳妥的写法是保留「在什么条件下」「在多大程度上」这类限定，让问题留出余地。
The IBO guide marks knowledge questions as contestable, centred on knowledge itself, and phrased in TOK concepts rather than subject terminology. A fourth condition sits underneath those three: the question has to stay open, with no correct answer hidden in the wording. Asking whether the scientific method is more reliable than the historical method sounds open but assumes reliability has one measure, and asking whether the criteria for reliability in different fields can be compared keeps the disagreement alive. Students often treat any yes or no question as open, when openness really means several defensible routes to different conclusions. Words such as most important or only usually announce that the answer is already inside the question, and building in conditions and degrees keeps the room to argue.
二次提问：从具体情境走到知识问题
考纲给的示范很直接。历史课上比较两位史学家对同一事件的不同判断，可以推到「专家的判断互相冲突时，我们凭什么在两者之间做选择」。一篇预测非洲未来人口增长的报纸文章，可以推到「一个明显不准确的模型为什么仍然有用」。一篇报道新药临床试验结果的期刊文章，可以推到「对这类研究应当设什么伦理约束」。这三个例子的共同动作是把情境退后一层，从「这件事是什么」转到「我们如何知道这件事」。学生在展评里最容易犯的错是停在第一层，用 950 词描述物件的来历，却没有把物件接到知识问题上。写展评时可以先写一句总结，说清这件物件让人看到哪个关于知识的问题，再看这句总结能不能对上 IA prompt 的原话。对不上就说明 prompt 选错了，或者物件选错了。
The guide's own examples show the move. Comparing two historians who judge the same event differently can lead to asking how we choose between expert judgements when they conflict, a newspaper piece predicting population growth in Africa can lead to asking why an obviously false model can still be useful, and a journal article on a drug trial can lead to asking what ethical constraints such research should carry. All three step back one level, from what happened to how we know it. The common failure in an exhibition is staying on that first level, spending the 950 words on where an object came from without connecting it to a knowledge question. One workable habit is writing a single sentence naming the knowledge issue the object exposes, then checking that sentence against the exact wording of the chosen prompt. If the two do not meet, either the prompt or the object is wrong.
论证骨架：主张、例子、推论、评估
TOK 的评分看论证，不看知识量。一个能拿到分的段落通常有四块：主张说清这一段要证明什么，例子给出具体的人、时间、地点或数据，推论解释例子怎样支持主张，评估说明这份支持有多强、在什么条件下会失效。缺任何一块都会掉分，缺推论最常见。学生写了一个很好的例子，然后直接跳到结论，中间那句「所以这个例子说明了什么」被省掉了。阅卷时这种段落读起来像资料汇编，因为读者得自己补上连接。把推论写出来需要一句明确的话，说明这个例子支持的是 A 而不是 B，理由是某个具体机制在起作用。评估则要承认反例存在，并交代为什么主张仍然成立。
TOK marks argument, not the volume of knowledge on display. A paragraph that earns credit usually carries four parts: a claim stating what this step proves, an example with real people, dates, places or data, a warrant explaining how the example supports the claim, and an evaluation of how far that support reaches. Leaving out the warrant is the most common gap, since students present a strong example and jump to the conclusion without the sentence that says what the example shows. In marking, such paragraphs read as a collection of material because the reader has to supply the link. A warrant has to be specific about mechanism: this example supports A rather than B because a particular process is doing the work. Evaluation then names a counter-case and explains why the claim survives it.
例子要落在真实情境里
展评的三件物件必须是真实存在的对象，或者是对象的照片，并且要交代它们的具体真实情境。一只旧手表可以用，前提是写清它属于谁、什么时候被使用、为什么与所选 prompt 有关。假想的物件不能用，泛泛的类别也不能用，「一本书」这样的说法不构成物件。论文里的例子同样要落到具体，写某一次具体的实验、某一篇具体的报道，比写「科学家们认为」有力得多。同班同学可以用同一条 IA prompt，但三件物件不能重复，考纲对这条写得很明确。常见的错误是选了很好看的东西，回头发现它在真实世界里与 prompt 只有很薄的联系，最后只能靠形容词撑着。
The three exhibition objects have to be real objects, or images of them, each identified with its specific real-world context. An old wristwatch works only if the commentary says whose it was, when it was used and why it bears on the chosen prompt, while invented objects and loose categories such as a book will not do. Examples in the essay need the same specificity, since one named experiment or one named news report carries more weight than a sentence beginning with scientists believe. Students in the same class may work from the same IA prompt but may not use any of the same objects, a rule the guide states plainly. A familiar mistake is choosing an attractive object that turns out to have a thin link to the prompt, which then has to be propped up with adjectives.
反方视角与让步：高分段的位置
高分段的 TOK 几乎都做过同一件事，把最强的反方说法写出来，再回应它。弱的反方长这样：有人认为不然，但这种看法不全面。这种写法既没写出对方的理由，也没承认对方哪里说得对。有效的让步有具体形状，先承认反方在某个范围内成立，再说明为什么在这个范围之外主张仍然站得住。写「数学证明不依赖观察」时，可以让步承认公理的选择受历史与文化影响，再说明公理一旦确定，推导过程不受观察左右。这种让步不会削弱论证，反而让论证显得知道自己站在哪里。评估部分不必给出最终裁决，TOK 接受「在多大程度上」这样的结论，只要理由链条完整。
High-scoring TOK writing almost always states the strongest opposing case and then answers it. The weak version is a sentence such as some people disagree, but that view is one-sided, which supplies neither the opponent's reason nor any concession. A useful concession has a shape: grant that the opposing case holds within a stated range, then show why the claim survives outside it. An essay arguing that mathematical proof does not rest on observation can concede that the choice of axioms carries historical and cultural weight, then argue that once axioms are fixed the derivation does not depend on observation. Such a concession makes an argument look located rather than weakened, and the evaluation need not deliver a verdict, since TOK accepts conclusions held to a degree as long as the chain of reasons is complete.
12 个概念是分析工具，不是装饰
IBO 列出的 12 个概念是 certainty、culture、evidence、explanation、interpretation、justification、objectivity、perspective、power、responsibility、truth、values。这些词不是贴在段首的标签，而是切开例子的刀。分析新闻核查的案例时，用 evidence 与 justification 可以分开两个问题，一个是有多少材料，一个是这些材料够不够支撑这一步推论。分析教科书争议时，用 power 与 perspective 可以问清谁有资格定稿，哪种叙述被挤出去了。学生常把概念名放在开头，后面接着写与这个概念无关的内容。更有效的做法是让概念出现在分析句里，例如指出这里的 justification 依赖同行评议，而同行评议在样本很小时最脆弱。一篇文章用透三个概念，比罗列十二个概念有用得多。
The twelve concepts the IBO lists are certainty, culture, evidence, explanation, interpretation, justification, objectivity, perspective, power, responsibility, truth and values. They work as tools for cutting into an example rather than as labels pasted at the start of a paragraph. In a fact-checking case, evidence and justification separate the question of how much material exists from the question of whether that material carries the inference. In a textbook dispute, power and perspective open up who was entitled to finalise the text and which account was pushed out. Concepts do their work inside analytical sentences, for instance noting that justification here rests on peer review and that peer review is weakest when samples are small. Three concepts used through an essay beat twelve concepts named and abandoned.
展评：三个物件对同一条 IA prompt
展评占 TOK 总成绩的 33%，满分 10 分，由任课老师评分，IBO 外部抽样复核。学生从 35 条 IA prompt 里选一条，题目的措辞一个字都不能改，三件物件都要接在这条 prompt 上。总字数上限 950 词，超出的部分阅卷人不再读，标题、图片里的文字、图注、参考文献与脚注不计入。考纲建议为这项任务投入约 8 小时教学时间。三件物件之间要能互相照亮，而不是三份互不相干的说明。写的时候可以让第一件立起一个角度，第二件给出反例，第三件收窄适用条件。丢掉最多分的地方是只解释物件为什么相关，没有说清它揭示了知识的哪个具体侧面。
The exhibition carries 33% of the TOK grade, is marked out of 10 by the teacher, and is moderated externally by the IBO. Students choose one of the 35 IA prompts, keep its wording unchanged, and link all three objects to that single prompt. The limit is 950 words, with examiners instructed to stop reading beyond it; the title, text inside images, captions, references and footnotes are excluded from the count. Around eight teaching hours are recommended for the task. Three objects should illuminate each other rather than sit as three separate notes, so one can open an angle, a second can supply a counter-case and a third can narrow the conditions. Marks are lost where a commentary explains only that an object is relevant without naming which aspect of knowledge it exposes.
论文：1600 词做两件事
论文占 67%，满分 10 分，由 IB 阅卷人评分。题目是 IB 每个考期发布的六道 prescribed titles，措辞不能改，题目在提交截止前约六个月挂在 programme resource centre 上。正文上限 1600 词，超过后阅卷人只读前 1600 词。这篇不是研究论文，但具体来源该标还是要标。选题时先看这道题要求比较哪两个 AOK，这一步能省掉大量废稿。1600 词做不了三件事，通常只够在两个领域各做一轮论证。学生最常见的结构问题是把题目拆成两半，前半篇写第一个领域，后半篇写第二个领域，两半之间没有比较。拿分的地方在比较句里，同一个方法在一个领域为什么有效，在另一个领域为什么失效。TK/PPF 要记录与老师的三次交流，这张表也得如实填。
The essay is worth 67%, is marked out of 10 by IB examiners, and must answer one of the six prescribed titles issued for the session without any change of wording; the titles appear on the programme resource centre about six months before the submission deadline. The limit is 1,600 words and examiners stop reading beyond it, while specific sources still need acknowledging even though the task is not a research paper. Reading the title for which areas of knowledge it asks you to compare saves a great deal of wasted drafting, since 1,600 words rarely supports more than one round of argument in each of two areas. The familiar structural failure splits the essay in half, one area first and the other second, with nothing between them. Marks live in the comparative sentence that explains why the same method works in one field and fails in another, and the TK/PPF form recording three conversations with the teacher has to be filled in honestly.
三类常见失分：描述、堆砌、复述结论
最常见的失分有三种。一种是描述代替论证，段落读完，读者知道了某个案子的来龙去脉，却不知道作者主张什么。另一种是例子堆砌，三个例子并排放着，之间没有递进，也没有一个被分析到底。还有一种出现在结尾，把开头的句子换个说法再写一遍，中间没有增加任何限定。这三类问题常常同时出现，因为根源相同，写作时想的是把知道的都摆出来，而不是把一句判断证明到位。改的办法很笨但管用，每段只留一个主张，例子只留最有力的那一个，写完问自己这一段到底证明了什么。删掉的内容往往比留下的更有说服力。
Three failures account for most lost marks. The first is description in place of argument, where a paragraph leaves the reader well informed about a case but unable to say what the writer claims. The second is a pile of examples standing side by side without progression, none of them analysed to the end. The third appears at the close, restating the opening sentence in new words without adding a single qualification. They usually arrive together because they share a cause, since the writer is trying to display what is known rather than prove one judgement. The fix is dull but effective: keep one claim and the strongest single example per paragraph, then ask what that paragraph proved. What gets deleted often carries more force than what stays.
67% 与 33%：时间怎么分
两项任务分值不同，投入的时间也该不同。论文 67%，展评 33%，但展评通常在第一年完成，论文在第二年。把展评拖到最后和大考撞在一起，是很多学校真实的节奏问题。展评的准备可以从日常物件开始，看到一件东西就问它让人对知识产生什么疑问，攒够五六个再回头对 35 条 prompt。论文的准备可以早一点读六道题，挑那道你能立刻想到两个领域例子的。IBO 对两项任务的建议教学时间是展评约 8 小时、论文约 10 小时，其余课堂时间用来谈概念与 AOK。字数上限不是目标，950 词与 1600 词的真正作用是让每一句都承重。
The two tasks carry different weights and deserve different amounts of time. The essay holds 67% and the exhibition 33%, yet the exhibition usually falls in the first year and the essay in the second, which is why pushing the exhibition back until it collides with exam season causes trouble in many schools. Exhibition preparation can start from ordinary objects, asking of each one what it makes you doubt about knowledge, then collecting five or six before turning back to the list of 35 prompts. Essay preparation benefits from reading the six titles early and choosing the one for which two areas of knowledge suggest examples at once. The IBO suggests roughly eight teaching hours for the exhibition and ten for the essay, with the rest of the course spent on concepts and areas of knowledge, and the word limits only work as a discipline that makes every sentence carry weight.
牛津高阶英汉双解词典 第10版 商务印书馆¥168📱 去京东购买京东精选
需要一对一辅导？Need One-on-One Tutoring?
A-Level / GCSE / IB / KS2-KS3 全科辅导
16621398022 同微信
Visit aleveler.com
或截图后用「扫一扫 → 相册」选择本图