[SSI Learning] 第十二期沙龙纪要：课堂科学论证的理论与实践（2021-06-02）
Minutes of the twelfth cloud salon:Tang Xiaowei’s understanding on in-class scientific argumentation (06/02/2021)

2021年6月2日19:30-21:00，北京师范大学社会性科学议题学习项目组在腾讯会议室举行在线沙龙，西南大学的唐小为教授进行《课堂科学论证的理论与实践》的专题报告。

一、课堂科学论证的概念

科学论证与科学辩论是紧密联系的一对概念：前者更强调产物，后者更强调过程；前者包含个体的论证，后者狭义地指代群体互动中的论证。这里不严格区分二者。最早关注课堂科学论证的是1999年Newton，Driver和Osborne，他们对英国多所中学的34堂科学课做课堂观察，考察科学论证在学校科学教学中的地位，认为科学论证在科学课堂中占比过低，这是现代科学课堂的重大缺陷。NGSS也认为这是科学教育中要加强的三大科学实践之一。我国课标没有单独提出科学论证，但也有相关表述，如“教师要为学生提供多样化的学习机会，如探究的机会、综合运用知识解决真实情境问题的机会、讨论辩论的机会、关心与环境、资源有关议题的机会等”。

科学论证以某种理论观点上的分歧为起点，持有不同观点的科学家为自己观点做辩护的过程称为科学辩论。科学研究中，科学家探究客观世界、不断产生新证据，不同观点的持有者都设法用自己的理论观点解释新发现的现象，其中解释力强的观点会暂时占上风，在一段时间内被当做普遍认可的“真理”。科学史就是一场旷日持久的科学辩论，更具有解释力的理论不断推翻旧的“真理”，更普适的理论涵盖特定情境的狭窄理论，从而推动科学不断向前发展。但科学家群体也与普通人一样具有思维定势，不愿意放弃旧有观点、可能迷信学术权威，通常新理论的证据要积累到相当程度，才会推翻既有的“真理”。

图1 科学史上的著名辩论——光的本质

Fig. 1: Famous debate in the history of science—Nature of Light

二、图尔敏论证结构模型(TAP)

图尔敏提出了一个论证结构模型，该模型认为，在一定证据的基础上产生某种结论，需要中介的推理过程，而这个推理过程反映了一定的理论视角，与此同时，也会考虑其他人对该推理的质疑，并就此进行补充说明（限定条件）。McNeill，Lizotte和Krajcik对图尔敏论证结构模型进行缩减，使其更便于应用，这一模型称之为CER，即Claim（直接回应问题的观点）、Evidence（证据）、Reasoning（连接证据与观点的推理）。Erduran,Simon和Osborne在2004年提出了辩论的五个层次，从低到高分别是：简单的两观点分歧；两观点有一定支持（证据、理论）但没有反驳；系列观点有一定支持且偶尔有较弱的反驳；多观点交互且有清晰反驳；多于一个反驳的拓展论证。以两段课堂上的师生对话展示CER模型、辩论五层次模型的应用。

图2 图尔敏论证结构模型

Fig. 2: Toulmin Argumentation Pattern

在使用CER模型进行结构分析的时候，应当注意以下几点：一，关注学生协调证据与理论的能力，学生的陈述中如果结构不全，应当追问其证据、推理；如果追问中学生仍在简单地复述观点而不能明确给出证据、推理，则应当帮助学生区分观点、证据和推理；如果存在与理论相左的证据，应当追问学生对该证据的看法，以考察学生协调证据与观点的能力；二，渐撤式提示，明确要求学生分开说明自己的观点、证据和推理，待学生逐渐习惯之后逐渐减少和撤去提示；三，有时难以明确区别一段话中的观点、证据、推理，这通常出现在论证结构复杂、情境中可以合理省略等情况下，虽然此时论证结构不明朗，但不能就此认为论证的质量低下。

三、作为TAP补充的其他论证模型

图尔敏模型有一些特定的缺陷，需要用其他模型进行补充。一，图尔敏模型不关注证据的认识层次，而这其实也应当成为论证质量考量的一环；二，图尔敏模型不关注证据质量差、证据与观点失配的问题，而这也会影响论证质量；三，图尔敏模型不关注辩论的目的，与辩论目的相符的论证才属于高质量辩论。

EBR模型解决了第一点。该模型认为根据观点到结论之间的连接形式，由低到高存在四个水平，即无支持（观点复述、循环论证）、现象推理（用数据支持）、关系推理（用表明局部规律关系的证据支持）、规则推理（用归纳法、演绎法的规则支持）。

对第二点的补充可以借鉴以下思路：在概念理解水平上，分为朴素概念、前科学概念、科学概念的三水平划分，这也可以作为科学论证质量考察的一个依据。

关于第三点的解决，有两个较为经典的模型。

一种模型认为，科学辩论有三种目标取向，即阐述想法（说明自己为什么有某个观点、认为一件事是怎么发生的）、意义建构（通过想法互动理解正在研究的现象是怎么回事）、说服他人（通过批判他人想法、为自己想法辩护，说服他人认同自己的想法），互动程度由低到高，价值也由低到高。Berland和Reiser认为“说服他人”取向由于违背社交的潜规则，在科学课堂中难以达成，但这又是最有价值的科学辩论形式，若想要在课堂中实现这一目标，要考虑以下几个因素：适于激发辩论的议题（给出相左的观点让学生选择并辩论，与学生前科学概念不同的观点、深入机制的开放性探究、SSI议题等）、互动板书（区别不同想法、整理双方思路）、物理设置（热席策略、“哲学家的椅子”、让观点双方面对面等）、反思回掷和传声筒（一方表述完毕后，询问另一方的想法；引导双方明确辩论发展的状况，使得辩论向更深层发展、更有针对性，避免同义反复等情况）、强调和表扬反驳。

另一种模型认为，在社会性科学议题中，探讨事物发生原理不再是主要目标，根据科学原理和社会需要进行社会决策才是关键，甚至价值观也可以作为论据，该理论认为论证分为三种，累积性论证（参与者仅分享各自观点，不评判他人观点）、争论性论证（参与者存在明确分歧，挑战他人观点的行为）、探索性论证（参与者基于彼此想法进行建构并达成共识），这三个目标取向与前述的三个目标取向一一对应，但该理论认为，议题本身具有很强的争议性，很容易形成分歧，分歧又很难达成共识，所以探索性论证比争论性论证更难实现，从决策导向的视角看探索性论证更有价值。

---

At 7:30-9:00 pm of June 2, 2021, the SSI Learning Project group of Beijing Normal University held a cloud salon in Tencent Conference Room. Professor Tang Xiaowei from Southwest University delivered a special report Theory and Practice of In-class Scientific Argumentation.

I. Concept of In-class Scientific Argumentation

Scientific argumentation and scientific debate are a pair of closely related concepts. The former emphasizes the outcome, while the latter emphasizes the process. The former includes the argumentation of an individual, while the latter refers to the argumentation in group interaction in a narrow sense. The two are not strictly distinguished here. Newton, Driver and Osborne were the first people to pay attention to in-class scientific argumentation in 1999. They observed 34 science classes at multiple middle schools in the U.K., to study the status of scientific argumentation in scientific teaching. They believe that the proportion of scientific argumentation in science classes is far lower than enough, which is a major shortcoming of modern science classes. NGSS also believes that it is one of the three major scientific practices that need to be addressed in science education. The curriculum in China does not propose scientific argumentation alone, but there are some relevant statements such as “teachers should provide students with various learning opportunities including the opportunities to explore, to solve real-life problems through comprehensive use of knowledge, to debate and discuss, and to care about issues related to environment and resources”.

Scientific argumentation starts with a certain theoretical point of view. The process in which scientists with different points of view defend their views is called a scientific debate. In scientific research, scientists explore the objective world, continuously produce new evidences, and people with different points of view try to explain their new findings with their own theories. Those who are good at persuasion will temporarily prevail and their opinions will be seen widely as the “truths” in a certain period. The history of science is a long scientific debate, where the more persuasive theories constantly overthrow the old “truths”, and the more general theories cover and replace the less general ones in a specific situation, thereby promoting the continuous scientific development. However, scientists are like ordinary people, who have similar mind set and will not give up their old points of view. Some probably are addicted to academic authority. Generally, evidences of new theories need to be accumulated to a certain amount before they can overthrow the existing “truths”.

II. Toulmin Argumentation Pattern (TAP)

Toulmin put forward a pattern of argumentation, which believes that if a certain conclusion is to be made based on a certain amount of evidence, it requires a deduction process that reflects a certain theoretical perspective. Meanwhile, it also considers others’ doubts of the deduction, and makes a supplementary explanation (with restrictions). McNeill，Lizotte and Krajcik simplified the model, so that it can be used more conveniently, and the simplified model is called CER, that is, Claim, Evidence, Reasoning. Erduran, Simon and Osborne put forward the five levels of a debate in 2004, that is, two simple viewpoints with differences, two viewpoints with a certain support (evidence, theories) but no objection; a series of viewpoints with a certain support but occasionally weak objection; multiple viewpoints that interact with each other with clear objections; and more than one rebutted extended argument. Let’s take two dialogues between students and teachers to show the application of the CER model and the five-level debate model.

When performing a structural analysis with CER model, we should pay attention to the following points. First, pay attention to students’ ability to coordinate evidence and theories. If the structure of the student’s statement is incomplete, we should ask about his/her evidence and deduction. If the student simply repeats his/her viewpoint rather than clarify his/her evidence and deduction, we should help the student to distinguish a viewpoint, evidence and deduction. If the evidence is inconsistent with the theory, we should ask the student his/her opinion about the evidence, in order to study his/her ability to coordinate evidence and theories. Second, to clearly ask students to distinguish a viewpoint, evidence and deduction; reduce or remove the hints when they get used to it. Third, sometimes it is hard to clearly distinguish a viewpoint, evidence and deduction in a dialogue, which is often seen in a complex argument structure or situation that can be reasonably omitted. Although this argument structure is not clear, the quality of the argumentation cannot be deemed as low.

III. Other Models of Argumentation as Supplements to the TAP

The Toulmin Argumentation Pattern has some flaws, which needs to be supplemented by other models of argumentation. First, the TAP does not pay attention to the knowledge level of the evidence, which in fact should be a part to evaluate the quality of the argument. Second, the TAP does not pay attention to problems such as low quality of evidence, and a mismatch between evidence and viewpoints, which will affect the quality of the argumentation. Third, the TAP does not pay attention to the purpose of the debate, but an argumentation that is not consistent with the purpose of the debate cannot be said to be a good debate.

The EBR model solves the first problem. This model believes that based on the connection from viewpoints to conclusions, there are four levels, that is, no support (repetitive viewpoint and circular argument), phenomenon reasoning (supported by data), relational reasoning (supported by evidence showing the relationship of local laws), and rule reasoning (supported by rules of induction and deduction).

The following thinking can be used as supplement to the second point. In terms of conceptual understanding, there are three levels, that is, simple concepts, pre-scientific concepts, and scientific concepts, which can also be used as a basis for the evaluation of the quality of scientific argumentation.

As for the third point, there are two typical models.

One model believes that there are three orientations of scientific debates, namely, explain ideas (explain why you have such a viewpoint and why you believe it happens in this way), construct meanings (understand what is going on through exchanges of ideas), and persuade others (persuade others into believing what you believe through judging their viewpoints and defending yours). The lower the level of interaction, the lower the value. Berland and Reiser believe that “persuading others” is against the unspoken rule of social interaction, so it is hard to be achieved in class. However, it is the most valuable form of scientific debates. If you want to achieve this in class, you need to consider the following factors: topics suitable for discussions (provide different viewpoints for students to choose and debate, viewpoints different from students’ pre-scientific concepts, in-depth open study of the mechanism, SSI, etc.), interactive traditional instruction (distinguish different ideas and sort out the thoughts of both sides), physical setting (hot seat strategy, “seat of the philosopher”, face-to-face debate, etc.), reflections and microphones (when one party finishes his part, ask about what the other party thinks; guide both parties to clarify the status of the debate, so that the debate can develop towards a deeper and more targeted way; avoid repetitive words), emphasize and compliment on objections.

The other one believes that in SSI, the discussion on how things happen is no longer a major goal. The key is to perform social decision-making based on scientific principles and social needs. Even values can be used as arguments, and this model believes that argumentation can be divided into three, namely, cumulative argumentation (the participants only share their own viewpoints without judging other’s), controversial argumentation (the participants obviously disagree with each other and challenge each other’s viewpoints), and exploratory argumentation (the participants construct their viewpoints and reach a consensus based on each other’s ideas). These three goal orientations are consistent with the three above-mentioned goals one by one. But the theory also believes that the topic itself is very controversial, which easily forms differences, and differences are hard to be reduced to reach a consensus. Therefore, exploratory argumentation is more difficult to be achieved than controversial argumentation, and it is more valuable to view exploratory argumentation from the perspective of decision orientation.