{ "schema_version": "v1", "id": "retrieve-learning-20260813", "date": "2026-08-13", "title": "Retrieve Learning：学习科学、古典音乐史与投资 conviction", "visibility": "public", "classification": { "category": "humanities-thought", "tags": [ "Retrieve Learning", "主动回忆", "学习科学", "古典音乐史", "投资写作", "conviction", "查漏补缺" ] }, "reading_source": { "label": "用户于 2026-08-13 进行的 Retrieve Learning：三项学习标的的无资料回忆", "provenance": "原始文字已完整保存到私有 本地原始材料（路径未公开）。本条 canonical entry 仅保留与三项学习标的直接相关的原始学习笔记；其他个人内容留在私有日记，不复制到 Git 或公开网页。" }, "raw_notes": [ "学习标的原始表述：包括《make it stick》，焦元溥的古典音乐历史，Aswath Damodaran关于Leo Schuchenner 的论文。", "Make it stick：整本书是基于近几十年来认知科学所研究的学习方法，给读者进行了介绍，包括好的方法和不好的方法，具有一定的论述性质，而不是仅仅把好的方法拿出来给人提供训练。\n(1)好的方法包括：通过提取学习来学习知识\n(2)不好的方法包括：反复阅读书本知识，死记硬背\n(3)我自己的心得笔记是直接摘录过来的：\nMake it stick is more likely an introduction of cognitive studies which show the good ways and bad ways of learning. However, as it writes in a scholarship way, it doesn’t explicitly say that which way is the best way and spends a big amount of space to argument why some ways are bad, such as repeating reading. For me, what I need to know is the best way and ready stick with it.\nMeanwhile, it seams that this book is not dedicated to learn writing. I found this book when I was reading Henry Winston’s book of how to speak. The ChatGPT tells me that make it stick is a similarly good book. But I don’t get it, I need to know what’s the value of reading this book, any point missing by me ?\nIn addition, I want to know what’s the latest common understanding in cognitive science, and what’s the most arguable point that people don’t understand each other. I want to achieve the latest level of cognitive science. I want to know the thesis, notion, scientists’ name, and study fields. I want to know if I need to explain the most important idea in cognitive science to the common people, what should I understand and how to speak it.\nAnother problem for make it stick is that how to learn things which are not totally correct but applied to certain regime, namely a blocked practice. How to differentiate the knowledge and information. How to really learning from this.", "古典音乐：\n(1)中古时期和巴洛克前中期的音乐不是重点\n(2)最早的音乐是跟教会相关，有记录的最早的音乐室欧xxx教皇汇编的乐谱，大概在隋朝时期，公元500年左右\n(3)到了公元1400年左右，音乐有了改观，从过去只有一个旋律的单调音乐，逐步增加旋律、增加每个音节之间的不同间隔数量，从最早8个音节或者4个音节，变化到8个、4个、3个、5个、6个等。\n(4)文艺复兴时期，音乐的中心是法国和英国，对文艺复兴中心意大利伦巴萨？家族的影响较小\n(5)到了巴洛克后期，音乐开始跟现代音乐接近了，有了更多层次的旋律，更多的变化", "Aswath的文笔很好，明显是以自己的口吻归纳了最近一段时间的故事，非常有功力。其中对于投资的conviction，是非常有意义的定义。在此基础上，我思考了用ai做交易的conviction，那就是用最强的模型和最强的推理深度，基于每天市场情况，从技术面角度来分析交易策略。我会为之执行一段时间，看看效果怎样。" ], "research_completions": [ { "question": "《Make It Stick》是否在主张存在一个脱离任务、学习者和反馈条件的唯一最佳学习方法？", "answer": "不是。它把检索练习、间隔、交错、生成、阐释和反馈组织成一套可组合的学习框架；反复阅读和集中练习的问题在于容易制造“看起来熟悉”的流畅感，不能单独代表长期学习。最新综述也把间隔和检索视为强证据方法，但强调学习者需要根据任务和延迟测试来调节，而不是机械执行一个万能配方。", "status": "verified", "learning_value": "你已经抓住了“提取优于被动重读”的核心，但需要把“最好的方法”改成“针对目标任务，能在延迟和新情境中成功提取并应用的组合”。", "source_ids": [ "makeitstick-official", "nature-effective-learning" ] }, { "question": "这段关于《Make It Stick》的英文心得如何改成自然、准确的英文？", "answer": "Make It Stick is best understood as an introduction to findings from cognitive science about effective and ineffective ways of learning. Because it presents evidence and explains why certain habits fail, it does not reduce learning to a single “best method” or a ready-made training program. Rereading can create a sense of familiarity without durable recall, whereas retrieval practice, spacing, interleaving, generation, and feedback work together to support longer-term learning.\n\nThe book is not a writing manual. I came across it while reading Henry Winston’s book on speaking, and ChatGPT recommended it as a related title. Its value for writing is indirect: after studying a writing principle, close the book, retrieve the principle, apply it in your own example, revisit it after a delay, and use feedback to revise. I should treat Make It Stick as a framework for designing and evaluating practice, not as a substitute for a book on rhetoric or composition.\n\nI also want a current map of cognitive science: its main claims, concepts, researchers, and fields; the areas where researchers disagree; and a plain-language explanation of the most important ideas. I especially want to understand how to learn conditional knowledge, where a rule works only under particular conditions. How should I distinguish information from knowledge, and how can I tell whether I have truly learned something?", "status": "paraphrase", "learning_value": "这版保留了你的问题意识，同时把“唯一最佳方法”“写作教材”和“条件性知识”分开表达；以后可把它作为英文复述的基准答案，再不看原文重写一遍。", "source_ids": [ "makeitstick-official", "retrievalpractice-guide" ] }, { "question": "这本书对学习写作有什么价值？如果它不是写作教材，为什么仍值得读？", "answer": "它的直接对象是学习和记忆，不是句法、篇章结构或修辞训练；因此它不能替代写作教材。它的价值在于提供一个“学习控制层”：读完写作书后先合上书，回忆规则、用自己的例子重写、间隔复习、交错比较不同文体，并在真实写作中获得反馈。你把它当作写作训练的操作系统，而不是写作内容本身，会更准确。", "status": "paraphrase", "learning_value": "你原先把“介绍学习科学”与“给我一本训练手册”作了二选一；遗漏点是这本书主要改变你选择练习、判断熟练度和安排复习的方式。", "source_ids": [ "makeitstick-official", "retrievalpractice-guide" ] }, { "question": "截至本次记录，认知科学关于学习的较稳固共识和仍在争论的部分是什么？", "answer": "较稳固的部分是：学习不能用即时熟悉感衡量；成功的检索、间隔安排、适当反馈和把知识用于目标任务，通常比单纯重读更能支持延迟保持。仍有边界的部分是：检索的难度、间隔长度、交错比例、先备知识和任务相似度如何组合；“学习”本身在不同学科仍没有一个完全统一的定义。近年的讨论特别强调训练后的提升究竟是任务特异的效率提升，还是能够迁移到未训练任务的更一般能力。", "status": "verified", "learning_value": "如果要向普通人解释，最稳妥的说法不是“某个技巧永远最好”，而是“用延迟后的无资料回忆检验自己，再根据错误反馈调整；是否迁移到新问题必须另测”。", "source_ids": [ "nature-effective-learning", "classroom-testing-meta", "learning-definition-2026", "transfer-review" ] }, { "question": "认知科学中最容易被双方说成互相矛盾、但其实需要区分的争论是什么？", "answer": "最典型的是“困难是否有益”。间隔、检索和交错可能让当下表现变差，却改善延迟保持；但困难不是越大越好。学习者若缺少足够先备知识、线索或反馈，困难会变成不可成功的负担。交错练习也不是无条件胜过 blocked practice：它最适合要求辨别类别、选择策略和应对变化的目标；初学者建立基本模式时，短期 blocked practice 可能更合适。", "status": "verified", "learning_value": "这正是你提出的“blocked practice”问题的关键：不要问哪一种练习绝对正确，而要问当前阶段的目标、先备知识、反馈和延迟测验是什么。", "source_ids": [ "desirable-difficulty", "interleaving-2024", "prior-knowledge" ] }, { "question": "blocked practice 是“只在某个 regime 下才正确的知识”吗？如何学习条件性知识？", "answer": "不是。blocked practice 指把同一类题目、案例或动作连续练习的编排方式；它不判断知识本身是否普遍正确。对于只在特定条件下成立的知识，应记录为“如果—那么—边界—反例—何时改用别的规则”，先用少量 blocked practice 建立规则和识别线索，再把相似规则交错，最后用未提示的案例检验能否选择正确规则。", "status": "paraphrase", "learning_value": "把“知识是否正确”和“练习顺序是否合适”分成两条轴，你就能避免把练习策略误当成真理判断。", "source_ids": [ "interleaving-2024", "desirable-difficulty", "prior-knowledge" ] }, { "question": "如何区分 information 和 knowledge，并判断自己是否真的学会？", "answer": "本条采用一个可操作而非哲学上唯一的区分：information 是外部可获得的事实、数据或说法；knowledge 是被组织进概念关系、能够在没有原文时被提取、能区分适用条件、能解释证据边界，并能在相近但不同的任务中使用的内容。判断标准不是“我看过”，而是能否无资料复述、举例和反例、说明何时不适用、接受反馈后修正，并在延迟后重新完成。", "status": "paraphrase", "learning_value": "你今后的笔记应同时记录事实、适用条件、证据等级和一个迁移任务，而不是只积累摘录。", "source_ids": [ "learning-definition-2026", "prior-knowledge", "transfer-review" ] }, { "question": "用户所说的“Leo Schuchenner”到底是谁？", "answer": "本次来源核验对应的人物是 Leopold（Leo）Aschenbrenner，而不是“Leo Schuchenner”。Damodaran 的原文还提到 Aschenbrenner 发布过 167 页的《Situational Awareness: The Decade Ahead》；Damodaran 的文章标题是《The Situational Awareness Blow-up: The Collateral Damage from Investing Conviction!》。", "status": "verified", "learning_value": "人名和作品名是检索入口，下一次 retrieve learning 遇到模糊专名时，应先标记为待核验，不要凭音近字补全。", "source_ids": [ "damodaran-conviction" ] }, { "question": "Damodaran 文中 conviction 的核心定义是什么？", "answer": "文中把 investment conviction 表述为：相信一项投资机会相对于其风险能够产生显著回报的程度，而且它是连续谱，不是只有“有信念/没信念”两个状态。 conviction 的形成至少涉及发现定价错误、相信市场会以某种方式修正、判断修正的时间范围，以及投资者自身的时间期限、性格和过度自信等特征。", "status": "verified", "learning_value": "你已经记住了“conviction 很有意义”，但下一步要能说出 conviction 的组成，而不把它简化为情绪强度。", "source_ids": [ "damodaran-conviction" ] }, { "question": "“用最强的模型和最强的推理深度”是不是 AI 交易的 conviction？", "answer": "这更像研究工具选择或研究流程承诺，还不是 conviction。若要形成可检验的 AI 交易 conviction，至少要写清：可重复的优势假设是什么、为什么技术面信号在该市场和时间尺度上应有优势、信号何时失效、回测和实时纸面记录如何避免前视偏差与数据泄漏、交易成本和容量如何处理、以及仓位和止损规则如何独立于“我很相信”。在证据不足前只能称为 hypothesis，不应称为 conviction，更不能直接产生实盘订单。", "status": "paraphrase", "learning_value": "把“对模型有信心”与“对可验证的市场优势有信念”分开；conviction 的对象应是可证伪的投资论证，而不是模型名气或推理长度。", "source_ids": [ "damodaran-conviction", "nature-effective-learning" ] }, { "question": "最早的音乐是否就是教会音乐，最早有记录的音乐是否约为公元 500 年、由教皇汇编？", "answer": "这句话把全球音乐史和西方教会记谱史混在了一起。人类最早的音乐无法直接确定，因为大量音乐没有留下记录；古代近东和希腊已有早于公元 500 年的音乐记谱遗存。若限定为西欧西方教会传统，plainchant 在加洛林文艺复兴后的 8—9 世纪开始被写下；所谓 Gregorian chant 与教皇 Gregory I 的联系带有传统归名，现存曲目和记谱传统并不能简单归为 Gregory I 在公元 590—604 年亲自汇编。因此“公元 500 年的教皇乐谱”应判为不准确。", "status": "incorrect", "learning_value": "以后先声明范围：全球最早音乐、古代最早记谱、最早西方教会记谱、Gregorian chant 的传统归名，分别是不同问题。", "source_ids": [ "oxford-plainchant", "oxford-earliest-notations", "cambridge-old-roman-chant" ] }, { "question": "公元 1400 年左右是否才从单旋律突然转向多旋律，并由 8、4 个音节扩展到 8、4、3、5、6 等间隔？", "answer": "这段回忆需要拆开并暂不接受其中的数字表述。西方复调的实践和理论在 9 世纪已有早期文献，12—13 世纪的巴黎圣母院传统已经发展出复杂复调；1400 年左右是文艺复兴早期的分期节点，不是复调的起点。你所说的“8、4、3、5、6”更像把八度、四度、三度、五度、六度等音程或协和观念与音节数量混在了一起；若没有焦元溥课程的具体页码或原句，不能把它恢复成确定的音乐史事实。", "status": "incorrect", "learning_value": "这是一条很好的测试题：回忆时不仅要写结论，还要保留术语类别、数量单位和证据位置；数字一旦无法解释，就标为待核验。", "source_ids": [ "oxford-earliest-notations", "met-renaissance" ] }, { "question": "文艺复兴音乐的中心是否主要是法国和英国，而意大利伦巴萨家族影响很小？", "answer": "这个判断过于二分。15 世纪早期确实有英国和北欧作曲家的重要影响，勃艮第宫廷尤其关键；约 1500 年，法兰德斯作曲家在欧洲艺术音乐中占据主导。与此同时，低地国家、法国和意大利之间存在密切交流，意大利的城市、宫廷、教会赞助和后来的 madrigal 等体裁都很重要。若你指的是 Medici（美第奇）家族，不能据此说意大利影响较小；更准确的说法是“作曲家网络具有跨国性，中心随时期、体裁和赞助制度移动”。", "status": "incorrect", "learning_value": "把“作曲技术中心”“赞助中心”“印刷中心”和“某一体裁的中心”分开，避免用一个国家替代整个欧洲网络。", "source_ids": [ "met-renaissance", "oxford-earliest-notations" ] }, { "question": "晚期巴洛克是否已经接近现代音乐，只是旋律层次更多、变化更多？", "answer": "可以把它作为听感入口，但不能当作完整历史判断。巴洛克常以复杂复调、装饰、对比、规范化体裁和功能调性为特征；晚期巴洛克确实建立了后来共同实践时期的重要语法，但古典时期仍在织体、句法、形式和乐器语言上发生明显变化。因此更准确的表述是：晚期巴洛克在调性和复杂组织上为后来的古典音乐铺路，但不是简单“越来越接近现代音乐”的直线进化。", "status": "paraphrase", "learning_value": "听音乐时同时问“旋律有几层”和“低音、和声、调性、句法、形式如何组织”，不要用层次多少概括整个时代。", "source_ids": [ "oxford-baroque" ] } ], "keywords": [ { "term": "retrieval practice", "definition": "在不看原文或只使用最少线索的情况下，从记忆中主动提取目标内容；随后用反馈校正。", "category": "学习科学", "usefulness": "把“我看过”转换成可观察的回忆能力，是本次 Retrieve Learning 的核心动作。", "source_ids": [ "makeitstick-official", "nature-effective-learning", "classroom-testing-meta" ] }, { "term": "spacing", "definition": "把学习或检索分散到有时间间隔的多个时点，而不是集中在一次完成。", "category": "学习科学", "usefulness": "让之后的提取不再依赖短期熟悉感，并支持延迟保持。", "source_ids": [ "nature-effective-learning" ] }, { "term": "interleaving", "definition": "交错安排不同但容易混淆的类别、问题类型或技能，使学习者需要辨别并选择策略。", "category": "学习科学", "usefulness": "适合训练“面对新题时先判断是哪一类问题”，而不只是重复同一种动作。", "source_ids": [ "interleaving-2024", "makeitstick-official" ] }, { "term": "blocked practice", "definition": "把同一类别或同一种动作连续安排在一起的练习顺序；它是编排方式，不是知识真假的判断。", "category": "学习科学", "usefulness": "适合在初期降低选择负担、建立基本模式，但不能替代后续的交错和迁移测试。", "source_ids": [ "interleaving-2024", "desirable-difficulty" ] }, { "term": "desirable difficulty", "definition": "一种会降低即时表现、但在学习者有足够线索和能力时改善长期保持或迁移的适度困难。", "category": "学习科学", "usefulness": "帮助解释为什么有效练习常常感觉不如重读顺畅，也提醒你不要把所有困难都美化成有益。", "source_ids": [ "desirable-difficulty", "nature-effective-learning" ] }, { "term": "fluency illusion", "definition": "因为材料读起来熟悉、流畅或刚刚看过，就高估自己之后能够独立回忆和应用的程度。", "category": "元认知", "usefulness": "是你想摆脱反复阅读时最需要监测的错觉。", "source_ids": [ "makeitstick-official", "nature-effective-learning" ] }, { "term": "transfer", "definition": "把在一个任务或情境中获得的知识、技能或策略用于新的任务或情境；近迁移和远迁移的距离不同。", "category": "学习科学", "usefulness": "检验你是否真的学会，而不是只会复述原文中的答案。", "source_ids": [ "transfer-review", "nature-effective-learning" ] }, { "term": "prior knowledge", "definition": "学习者已经组织在记忆中的相关知识；它必须被激活、与当前任务相关且与新内容相容，才稳定地帮助学习。", "category": "学习科学", "usefulness": "解释为什么同一种困难对初学者可能过重，对有基础者却可能正合适。", "source_ids": [ "prior-knowledge" ] }, { "term": "investment conviction", "definition": "相信一项投资机会相对于其风险能够产生显著回报的信念程度；它是连续谱。", "category": "投资推理", "usefulness": "把“我很看好”拆成定价错误、修正机制、时间范围和投资者特征。", "source_ids": [ "damodaran-conviction" ] }, { "term": "mispricing", "definition": "资产市场价格与投资者认为的合理价值或未来现金流价值之间的定价偏差。", "category": "投资推理", "usefulness": "conviction 的起点必须是可说明的机会判断，而不是模型强或观点强。", "source_ids": [ "damodaran-conviction" ] }, { "term": "market correction", "definition": "市场价格偏差通过信息、事件、催化剂、期限到达或其他机制被重新定价的过程。", "category": "投资推理", "usefulness": "没有修正机制和时间范围，发现“便宜”不等于有 conviction。", "source_ids": [ "damodaran-conviction" ] }, { "term": "plainchant", "definition": "西方教会传统中的单声部礼仪歌唱类型；现存西方书写传统主要在 8—9 世纪以后逐渐形成。", "category": "音乐史", "usefulness": "建立教会音乐、Gregorian chant 和后续复调发展之间的时间线。", "source_ids": [ "oxford-plainchant", "cambridge-old-roman-chant" ] }, { "term": "neumatic notation", "definition": "早期西方记谱中的纽姆符号体系，最初主要帮助歌者回忆旋律走向，并不等于现代五线谱的完整精确记录。", "category": "音乐史", "usefulness": "纠正“公元 500 年教皇已经汇编出现代意义乐谱”的时间混淆。", "source_ids": [ "oxford-plainchant", "oxford-earliest-notations" ] }, { "term": "polyphony", "definition": "多个相对独立的旋律线同时进行的复调织体。", "category": "音乐史", "usefulness": "把“旋律层次变多”转换为可听、可分析的织体概念，并放回 9—13 世纪以来的连续发展。", "source_ids": [ "oxford-earliest-notations", "met-renaissance" ] }, { "term": "Franco-Flemish", "definition": "通常指 15—16 世纪来自低地国家及法语区、在欧洲各地活动的作曲家网络与风格传统。", "category": "音乐史", "usefulness": "比“文艺复兴中心是法国/英国”更准确地描述约 1500 年欧洲艺术音乐的跨国中心。", "source_ids": [ "met-renaissance" ] }, { "term": "functional tonality", "definition": "以主音、属音等和声功能及其方向性组织调性关系的体系；在巴洛克时期逐步成熟。", "category": "音乐史", "usefulness": "解释晚期巴洛克为何在和声组织上连接到后来的共同实践时期，而不只是“旋律更多”。", "source_ids": [ "oxford-baroque" ] } ], "vocabulary": [ { "id": "retrieval-practice", "term": "retrieval practice", "translation": "检索练习；提取练习", "language": "en", "part_of_speech": "noun phrase", "definition": "Actively recalling learned material from memory, often followed by corrective feedback.", "category": "学习科学", "levels": [ "CEFR-C1", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "学习科学研究中的核心术语，抽象度和专业搭配较高。", "example": "Retrieval practice is more than rereading a familiar page.", "notes": "retrieve learning 是较自然的口语表达；正式术语通常写 retrieval practice。", "source_context": "本次记录的核心学习方法。", "source_locator": "用户 Retrieve Learning 说明；学习科学综述", "status": "verified", "source_ids": [ "nature-effective-learning", "classroom-testing-meta" ] }, { "id": "spacing", "term": "spacing", "translation": "间隔学习；分散练习", "language": "en", "part_of_speech": "noun", "definition": "Distributing study or practice across time rather than massing it in one session.", "category": "学习科学", "levels": [ "CEFR-B2", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "medium", "level_rationale": "普通词在学习科学中有专门搭配 spacing effect。", "example": "Spacing makes later retrieval less dependent on short-term familiarity.", "source_context": "0/1/3/7/14/21 日复习安排的理论基础。", "source_locator": "学习科学综述", "status": "verified", "source_ids": [ "nature-effective-learning" ] }, { "id": "interleaving", "term": "interleaving", "translation": "交错练习；交错学习", "language": "en", "part_of_speech": "noun / gerund", "definition": "Alternating between related categories or problem types so that the learner must discriminate and select a response.", "category": "学习科学", "levels": [ "CEFR-C1", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "教育心理学和学习科学中的专业术语。", "example": "Interleaving is useful when the learner must decide which rule applies.", "source_context": "与 blocked practice 对照的练习编排。", "source_locator": "npj Science of Learning 2024", "status": "verified", "source_ids": [ "interleaving-2024" ] }, { "id": "blocked-practice", "term": "blocked practice", "translation": "集中式练习；分块练习", "language": "en", "part_of_speech": "noun phrase", "definition": "Practicing one category or type consecutively before moving to another.", "category": "学习科学", "levels": [ "CEFR-C1", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "专门术语，不能按普通 block 的字面义理解。", "example": "Blocked practice can help establish a basic pattern before interleaving.", "source_context": "用户对“某一 regime 下适用”的困惑。", "source_locator": "npj Science of Learning 2024", "status": "verified", "source_ids": [ "interleaving-2024" ] }, { "id": "desirable-difficulty", "term": "desirable difficulty", "translation": "有益困难；适宜困难", "language": "en", "part_of_speech": "noun phrase", "definition": "A learning challenge that hurts immediate performance but can improve later retention when it remains achievable.", "category": "学习科学", "levels": [ "CEFR-C1", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "认知心理学专门概念，要求理解条件限制。", "example": "A desirable difficulty is not the same as making a task impossible.", "source_context": "解释为什么提取比重读更费力。", "source_locator": "Health professions education review", "status": "verified", "source_ids": [ "desirable-difficulty" ] }, { "id": "transfer", "term": "transfer", "translation": "迁移", "language": "en", "part_of_speech": "noun", "definition": "Applying learning from one task or context to another task or context.", "category": "学习科学", "levels": [ "CEFR-B2", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "普通词在认知科学中有 near/far transfer 的专门用法。", "example": "A delayed transfer test asks whether the learner can use the idea in a new case.", "source_context": "本次将“会复述”与“会应用”区分开。", "source_locator": "Nature Reviews Psychology transfer review", "status": "verified", "source_ids": [ "transfer-review" ] }, { "id": "conviction", "term": "conviction", "translation": "确信；信念强度；投资信念", "language": "en", "part_of_speech": "noun", "definition": "In Damodaran’s investing context, the degree of belief that an opportunity can generate significant returns relative to its risks.", "category": "投资推理", "levels": [ "CEFR-B2", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "日常词在投资语境中承担结构化判断含义。", "example": "Conviction should be tied to an edge, a correction mechanism, and a time horizon.", "source_context": "Damodaran 对投资 conviction 的定义与后果分析。", "source_locator": "Damodaran original post", "status": "verified", "source_ids": [ "damodaran-conviction" ] }, { "id": "mispricing", "term": "mispricing", "translation": "错误定价；定价偏差", "language": "en", "part_of_speech": "noun", "definition": "A market price believed to differ from an asset’s reasonable or intrinsic value.", "category": "投资推理", "levels": [ "CEFR-C1", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "估值与资产定价中的专业派生词。", "example": "Spotting mispricing is not enough without a plausible correction path.", "source_context": "conviction 的起点。", "source_locator": "Damodaran original post", "status": "verified", "source_ids": [ "damodaran-conviction" ] }, { "id": "plainchant", "term": "plainchant", "translation": "素歌；单声部教会歌唱", "language": "en", "part_of_speech": "noun", "definition": "A broad term for unaccompanied monophonic liturgical chant in the Western church tradition.", "category": "音乐史", "levels": [ "CEFR-C1", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "音乐史专门名词，不能与所有宗教音乐等同。", "example": "Western plainchant was written down in the wake of the Carolingian renaissance.", "source_context": "纠正“最早音乐就是教会音乐”的范围混淆。", "source_locator": "Oxford University Press, Western Plainchant", "status": "verified", "source_ids": [ "oxford-plainchant" ] }, { "id": "neumatic-notation", "term": "neumatic notation", "translation": "纽姆记谱法", "language": "en", "part_of_speech": "noun phrase", "definition": "An early Western notation system using neumes, initially functioning mainly as mnemonic guidance for chant.", "category": "音乐史", "levels": [ "CEFR-C1", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "历史音乐学术语。", "example": "Neumatic notation did not begin as a complete modern staff notation.", "source_context": "教会记谱史的关键纠错点。", "source_locator": "Oxford and Cambridge music-history sources", "status": "verified", "source_ids": [ "oxford-plainchant", "oxford-earliest-notations" ] }, { "id": "polyphony", "term": "polyphony", "translation": "复调", "language": "en", "part_of_speech": "noun", "definition": "A texture in which multiple relatively independent melodic lines sound simultaneously.", "category": "音乐史", "levels": [ "CEFR-C1", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "音乐理论核心术语。", "example": "Renaissance polyphony developed through several earlier medieval practices.", "source_context": "修正“1400 年才从单旋律突然变复调”。", "source_locator": "Oxford History of Western Music; Met Renaissance essay", "status": "verified", "source_ids": [ "oxford-earliest-notations", "met-renaissance" ] }, { "id": "functional-tonality", "term": "functional tonality", "translation": "功能调性", "language": "en", "part_of_speech": "noun phrase", "definition": "A tonal system organized by directional harmonic functions such as tonic and dominant.", "category": "音乐史", "levels": [ "CEFR-C1", "Academic/PhD" ], "level_basis": "expert_estimate", "level_confidence": "high", "level_rationale": "和声与音乐史中的专业概念。", "example": "Functional tonality emerged during the Baroque period and shaped later common-practice music.", "source_context": "把晚期巴洛克与后续古典音乐的联系具体化。", "source_locator": "Oxford Bibliographies in Music", "status": "verified", "source_ids": [ "oxford-baroque" ] } ], "reading_contexts": [ { "id": "learning-science-context", "summary": "本次对《Make It Stick》的回忆准确抓住了检索练习、反复阅读的局限和“感觉困难但长期更有效”的直觉冲突。需要补全的是：检索应与反馈、间隔、交错、先备知识和目标任务结合；学习是否发生要用延迟保持和迁移测试确认，而不是以当下顺畅为标准。", "author": "Peter C. Brown、Henry L. Roediger III、Mark A. McDaniel；相关研究综述作者包括 Shana K. Carpenter、Steven C. Pan、Andrew C. Butler", "work": "Make It Stick: The Science of Successful Learning", "era": "21 世纪学习科学与认知心理学", "people": [ "Henry L. Roediger III", "Jeffrey D. Karpicke", "Robert A. Bjork", "Elizabeth L. Bjork", "Shana K. Carpenter", "Steven C. Pan", "Andrew C. Butler" ], "writing_style": "以研究、案例和反直觉结果论证学习方法，不把复杂证据压缩成单一口号。", "writing_use": "写作训练时可把每次阅读转为无资料提纲、例文重写、延迟复述和针对真实读者的迁移任务；写作本身仍需另配修辞、结构和句法训练。", "register": "学习科学科普、研究转译、教育心理学", "collocations": [ "retrieval practice", "spaced practice", "interleaved practice", "desirable difficulty", "delayed retention", "transfer test" ], "source_context": "本次用户自述与学习科学综述的对照。", "source_locator": "Make It Stick 全书主题；2022 Nature Reviews Psychology；2021 classroom meta-analysis", "source_ids": [ "makeitstick-official", "nature-effective-learning", "classroom-testing-meta", "interleaving-2024" ] }, { "id": "music-history-context", "summary": "你已经记住“教会—记谱—复调—文艺复兴—巴洛克”的大方向，但把全球最早音乐、西方教会音乐的书写起点、Gregorian chant 的传统归名、复调的发展和文艺复兴的地域网络压缩成了一条错误的直线。更稳妥的骨架是：古代已有记谱遗存；西方 plainchant 在 8—9 世纪以后逐步书写；复调在中世纪已发展；15 世纪早期有英国与北欧、勃艮第网络，约 1500 年法兰德斯作曲家影响突出，同时意大利的赞助和体裁发展不可忽略；巴洛克逐步成熟功能调性并为古典时期铺路。", "author": "Richard Taruskin、David Hiley；大都会艺术博物馆教育部门的 Rebecca Arkenberg", "work": "西方音乐史：从最早记谱、plainchant 到文艺复兴和巴洛克", "era": "古代记谱遗存；西欧中世纪（约 8—15 世纪）；文艺复兴（约 1400—1600）；巴洛克（约 17 世纪至 18 世纪前半）", "cultural_background": "教会、宫廷、城市赞助、印刷和跨地区作曲家网络共同决定了音乐作品的保存与传播；“中心”应随时期、体裁和机构变化。", "people": [ "Gregory I", "Guillaume Du Fay", "Josquin des Prez", "Claudio Monteverdi", "J. S. Bach" ], "writing_style": "音乐史需要用时间线、织体、记谱、赞助和体裁五条线并行，而不能只用“越来越复杂”概括。", "writing_use": "下一次复述先画四列：年代、主要织体、记谱/传播、中心机构；每一项只写有证据的词。", "register": "音乐史、音乐理论、文化史", "collocations": [ "Western plainchant", "Gregorian chant", "neumatic notation", "early polyphony", "Franco-Flemish composers", "functional tonality" ], "source_context": "用户对焦元溥古典音乐历史学习内容的回忆与外部音乐史资料对照。", "source_locator": "Oxford History of Western Music；Oxford Western Plainchant；Met Music in the Renaissance", "source_ids": [ "oxford-plainchant", "oxford-earliest-notations", "cambridge-old-roman-chant", "met-renaissance", "oxford-baroque" ] }, { "id": "damodaran-conviction-context", "summary": "你对 Damodaran 文章的阅读理解有两个强点：感受到作者用自己的口吻组织近期故事，并抓住 conviction 不是简单褒义词。需要补全的是，文章讨论的不是“谁使用最强 AI 模型”，而是投资者如何从定价错误、市场修正、时间范围和自身特征形成信念，并把 conviction 转化为集中度与杠杆。把这个框架迁移到 AI 交易时，应先定义可证伪的 edge，再记录反证、成本、容量、失效条件和风险预算。", "author": "Aswath Damodaran", "work": "The Situational Awareness Blow-up: The Collateral Damage from Investing Conviction!", "era": "2026 年投资评论与 AI 投资叙事", "people": [ "Aswath Damodaran", "Leopold Aschenbrenner" ], "writing_style": "先讲近期案例，再将案例抽象成定义、维度和后果；同时保留 caveat、运气、杠杆、费用和清算等限制。", "writing_use": "可学习“案例—定义—机制—后果—限定条件”的投资论证结构，尤其不要只写回报而不写风险来源。", "register": "财经评论、估值与投资推理", "collocations": [ "investing conviction", "relative to its risks", "market correction", "time horizon", "position sizing", "financial leverage" ], "source_context": "Damodaran 原文对 Leopold Aschenbrenner、Situational Awareness 基金和 conviction 的讨论。", "source_locator": "Damodaran original post", "source_ids": [ "damodaran-conviction" ] } ], "quotes": [], "reading_comprehension": { "summary": "本次 Retrieve Learning 的核心成果不是把三份材料复述得更漂亮，而是发现三种不同的知识结构：学习科学要记“机制—边界—测试”；音乐史要记“时间线—织体—机构网络”；投资文章要记“定义—形成机制—行动后果—风险限定”。你对《Make It Stick》的主旨和 Damodaran 的 conviction 已有可用骨架；最大缺口集中在精确术语、时间线、证据层级，以及把个人判断直接升级为一般规律。", "concept_connections": [ "检索练习要求把内容从记忆中提取出来；音乐史和 conviction 的复述也应采用同一动作，而不是只重读讲义。", "blocked practice 与 interleaving 是练习顺序，不是知识真假的二元判断；条件性知识需要额外记录适用边界和反例。", "音乐史中的“中心”与投资中的“优势”都不能只看一个漂亮结论：前者要看作曲家、机构和传播，后者要看定价错误、修正机制、期限和风险。", "Damodaran 文章的“conviction—集中—杠杆”链条可以作为 AI 交易研究的风险审查框架，但不能把模型能力当成已验证的市场优势。" ], "market_implications": [ "对 AI 交易而言，“模型强、推理深、每天分析”只能描述工具和流程；要成为可检验论证，必须另行记录 edge、时间尺度、反证、交易成本、容量、数据泄漏和失效条件。", "阅读投资案例时，将描述性叙事、作者的解释和可验证的投资结论分栏；高回报、名人支持和强 conviction 本身都不是可持续优势的充分证据。", "本条只提供学习与研究设计，不输出自动订单、收益承诺或仓位建议。" ], "implications_label": "学习科学、音乐史与投资推理" }, "market_update": { "as_of": "2026-08-13", "section_title": "学习研究与 AI 交易应用更新", "summary": "截至本次记录，公开研究支持检索和间隔作为稳固的学习工具，但并不支持把任何单一技巧说成脱离目标的万能方案。当前最值得保留的争论是难度、先备知识和迁移边界如何校准；这也决定了 AI 交易 conviction 只能通过预先定义的研究记录和延迟验证逐步形成。", "points": [ { "topic": "检索与间隔", "current_state": "学习科学综述和课堂元分析支持测试/检索对长期学习有帮助，但效果会受到控制条件、反馈、重复次数、材料匹配和测试时间点调节。", "implication": "每日训练要保留错误和反馈，不要只记录答对次数。", "source_ids": [ "nature-effective-learning", "classroom-testing-meta" ] }, { "topic": "交错、困难与迁移", "current_state": "交错可促进比较和策略选择，但在真实自组织学习中仍需结合类别相似度、先备知识和目标；任务上的提升也不自动等于远迁移。", "implication": "每周至少加入一次未提示的新案例，并把“能否迁移”单独评分。", "source_ids": [ "interleaving-2024", "desirable-difficulty", "transfer-review", "prior-knowledge" ] }, { "topic": "AI 交易 conviction", "current_state": "本次只形成研究假设，没有形成经过校准的交易优势证据。", "implication": "先做历史回放和纸面记录；预注册信号、基准、成本、容量和失效规则，经过延迟评估后再讨论是否值得承担任何真实风险。", "source_ids": [ "damodaran-conviction" ] } ] }, "knowledge_updates": [ { "original_idea": "最好的学习方法就是提取学习，其他方法都不重要。", "verdict": "需要补全", "updated_view": "检索是核心动作，但要与初始理解、适度 blocked practice、间隔、交错、反馈和迁移任务组合；方法的好坏取决于学习目标和阶段。", "practical_change": "每次读完先做无资料回忆，再根据错误决定是补读、分块练习还是交错练习。" }, { "original_idea": "重复阅读和死记硬背是完全没有价值的坏方法。", "verdict": "过度绝对化", "updated_view": "把重复阅读当成唯一方法会制造流畅感；短暂重读仍可用于建立初始理解、定位遗漏和检查反馈，但必须让位给独立提取和应用。", "practical_change": "重读前先写出已知内容，重读后立刻合上材料复述，并留下一个新情境测试。" }, { "original_idea": "blocked practice 是某种只在特定 regime 下才正确的知识。", "verdict": "概念混淆", "updated_view": "blocked practice 是练习顺序；regime 条件属于知识的适用边界。两者要分别建模。", "practical_change": "每个条件性规则用“如果—那么—边界—反例—切换条件”记录，再安排交错案例。" }, { "original_idea": "AI 交易的 conviction 等于使用最强模型和最强推理深度。", "verdict": "尚未成立", "updated_view": "模型和推理深度是研究工具变量；conviction 的对象应是可证伪的市场优势及其修正机制，且必须受成本、容量、失效和风险约束。", "practical_change": "把这项计划标记为 hypothesis，先用历史回放、纸面交易和预注册规则检验，不直接下实盘订单。" }, { "original_idea": "最早有记录的音乐是公元 500 年左右教皇汇编的教会乐谱，1400 年左右才出现复调。", "verdict": "事实错误与时间线压缩", "updated_view": "古代已有更早的音乐记谱遗存；西欧 plainchant 的书写传统主要在 8—9 世纪以后形成；复调在中世纪已发展，1400 年是文艺复兴早期分期节点而非复调起点。", "practical_change": "用全球遗存、西方教会记谱、复调发展、文艺复兴网络四条时间线重画笔记。" }, { "original_idea": "文艺复兴音乐的中心是法国和英国，意大利影响较小。", "verdict": "过度二分", "updated_view": "早期有英国和北欧、勃艮第网络的突出影响；约 1500 年法兰德斯作曲家占据主导，但意大利的赞助、城市和体裁发展与跨国作曲家流动同样重要。", "practical_change": "复述时按时期、体裁和机构说“中心”，不再为整个时代指定一个国家。" } ], "next_actions": [ { "action": "每天用 8—12 分钟做一次无资料 Retrieve Learning：写出当天学到的 3 个概念、1 个边界、1 个例子和 1 个仍不确定的问题。", "why": "固定格式会把回忆从泛泛感想变成可评分的提取任务。", "evidence_required": "不看材料完成后，再用来源核对并标记 verified、incorrect 或 unverified。", "guardrail": "不要把一次顺利复述当成长期掌握；至少保留一个延迟复测日期。" }, { "action": "按 0/1/3/7/14/21 日节奏复习本条测试卡；每次只先看问题，再写答案，最后看来源和纠错。", "why": "系统已有透明的间隔复习队列，适合把今天的回忆变成可追踪的长期记忆。", "evidence_required": "记录每张卡的答对、部分正确、错误和修正，而不是只记录是否看过。", "guardrail": "21 天是复习节奏，不是“21 天养成习惯”的科学结论。" }, { "action": "每周做一次交错测试：从学习科学、音乐史、Damodaran 三个主题各抽两题，先判断问题类型，再回答。", "why": "你目前的主要缺口是把概念、时间线和投资框架混成顺口叙述；跨主题交错能检验是否真正会辨别。", "evidence_required": "至少有一题要求解释适用边界或反例，不能只问定义。", "guardrail": "初学新概念时可以先短暂 blocked practice；达到基本准确后再交错。" }, { "action": "把 AI 交易想法写成一页预注册研究卡：edge、市场/时间尺度、信号、基准、数据截止点、成本、容量、失效条件、评价指标和复盘日期。", "why": "把“最强模型”从身份性信念改成可证伪的研究假设，才能判断 conviction 是否有证据。", "evidence_required": "使用真实历史数据或真实纸面记录，并保留命令、数据源、日志和结果；先做延迟验证。", "guardrail": "未完成校准和风险审查前，不自动下单、不承诺收益、不因 conviction 提高杠杆。" }, { "action": "为焦元溥的下一段音乐史材料建立四列表：年代、织体/和声、传播与赞助、代表人物或体裁；具体作品再进入独立音乐记忆库。", "why": "本次没有提供具体作品、作曲家录音或听感，暂不虚构 music-library 记录；先把历史框架补稳。", "evidence_required": "下一次至少提供一个具体作品或录音，才能做旋律锚点和听—查—再听卡。", "guardrail": "不把无法定位的“8、4、3、5、6”数字写成已核验音乐史事实。" } ], "risk_guardrails": [ "本条的认知科学总结不是“最新全部研究”的穷尽清单；“学习”在跨学科研究中仍有定义差异，来源和证据状态要随新材料更新。", "不要把即时流畅、记得一句话或模型给出漂亮解释当成学习完成；必须加入延迟提取、反例和迁移测试。", "不要把 desirable difficulty 理解成越难越好；没有足够先备知识、线索或反馈时，困难可能变成无效负担。", "不要把 Damodaran 的投资评论、历史回报叙事或自己的 conviction 当成经过校准的交易优势。", "AI 交易研究先使用真实历史回放或纸面记录；真实资金、杠杆、自动订单和仓位大小不由本读书笔记生成。", "公开网页只展示整理后的学习内容；原始日记文字、私人经历和内部路径保留在私有日记或本地审计层。" ], "review_cards": [ { "id": "retrieve-learning-20260813-cognitive-best-method", "kind": "concept", "prompt": "为什么不能把 retrieval practice 简化为脱离任务的“唯一最佳学习方法”？", "answer": "因为学习效果取决于目标、先备知识、反馈、间隔、交错和延迟迁移；检索是核心动作，但不是独立万能配方。", "source_ids": [ "nature-effective-learning", "makeitstick-official" ] }, { "id": "retrieve-learning-20260813-cognitive-fluency", "kind": "concept", "prompt": "反复阅读最危险的地方是什么？", "answer": "它容易提高熟悉感和即时流畅度，让人误以为之后可以独立回忆和应用；应先无资料提取，再用反馈检查。", "source_ids": [ "makeitstick-official", "nature-effective-learning" ] }, { "id": "retrieve-learning-20260813-cognitive-blocked", "kind": "concept", "prompt": "blocked practice 与“只在某个 regime 下才正确的知识”有什么区别？", "answer": "blocked practice 是把同一类材料连续练习的顺序；regime 是知识的适用条件。前者属于学习编排，后者属于知识内容和边界。", "source_ids": [ "interleaving-2024", "desirable-difficulty" ] }, { "id": "retrieve-learning-20260813-cognitive-difficulty", "kind": "concept", "prompt": "什么时候困难是 desirable，什么时候会变成 undesirable？", "answer": "当学习者有足够先备知识和线索、仍能成功完成检索且困难改善延迟保持时是 desirable；若几乎无法成功、没有反馈或超过当前能力，困难就可能妨碍学习。", "source_ids": [ "desirable-difficulty", "prior-knowledge" ] }, { "id": "retrieve-learning-20260813-cognitive-transfer", "kind": "concept", "prompt": "如何证明自己不是只记住了信息，而是形成了知识？", "answer": "无资料复述核心关系，说明适用条件和反例，接受反馈后修正，并在延迟后的相似但不同案例中选择和解释正确规则。", "source_ids": [ "transfer-review", "learning-definition-2026" ] }, { "id": "retrieve-learning-20260813-music-earliest", "kind": "concept", "prompt": "请区分：最早音乐、最早古代记谱、最早西方教会记谱、Gregorian chant 的传统归名。", "answer": "最早音乐无法直接确定；古代近东和希腊已有早于公元 500 年的记谱遗存；西方 plainchant 的书写传统主要在 8—9 世纪以后形成；Gregorian chant 与 Gregory I 的关系主要是传统归名，不能等同于公元 500 年已有现存现代乐谱。", "source_ids": [ "oxford-plainchant", "oxford-earliest-notations", "cambridge-old-roman-chant" ] }, { "id": "retrieve-learning-20260813-music-polyphony", "kind": "concept", "prompt": "复调是不是在公元 1400 年左右才出现？", "answer": "不是。早期复调的理论和实践在 9 世纪已有文献，12—13 世纪已有复杂发展；1400 年左右是文艺复兴早期的分期节点。", "source_ids": [ "oxford-earliest-notations" ] }, { "id": "retrieve-learning-20260813-music-centers", "kind": "concept", "prompt": "如何更准确地描述文艺复兴音乐的“中心”？", "answer": "15 世纪早期英国和北欧、勃艮第网络重要；约 1500 年法兰德斯作曲家影响突出；意大利的赞助、城市和体裁发展又与跨国作曲家网络相互作用，不能简单说意大利影响很小。", "source_ids": [ "met-renaissance" ] }, { "id": "retrieve-learning-20260813-music-baroque", "kind": "concept", "prompt": "晚期巴洛克与后来的古典音乐是什么关系？", "answer": "巴洛克逐步成熟功能调性、复杂复调、装饰和规范化体裁，为共同实践时期提供重要语法；但古典时期仍改变织体、句法、形式和乐器语言，不能说只是线性地“更接近现代”。", "source_ids": [ "oxford-baroque" ] }, { "id": "retrieve-learning-20260813-damodaran-name", "kind": "concept", "prompt": "Damodaran 文章中的 Leo 全名和他发布的论文是什么？", "answer": "Leopold（Leo）Aschenbrenner；文章提到他的《Situational Awareness: The Decade Ahead》。", "source_ids": [ "damodaran-conviction" ] }, { "id": "retrieve-learning-20260813-damodaran-definition", "kind": "concept", "prompt": "Damodaran 如何定义 investment conviction？", "answer": "相信一项投资机会相对于其风险能够产生显著回报的程度；它是连续谱，形成还涉及定价错误、市场修正、时间范围和投资者特征。", "source_ids": [ "damodaran-conviction" ] }, { "id": "retrieve-learning-20260813-damodaran-consequences", "kind": "concept", "prompt": "为什么 conviction 会影响投资结果？", "answer": "它会影响仓位集中度和使用金融杠杆的意愿，因此信念强度不只是心理状态，也会改变风险暴露和损失路径。", "source_ids": [ "damodaran-conviction" ] }, { "id": "retrieve-learning-20260813-ai-conviction", "kind": "application", "prompt": "“最强模型+最深推理”还缺什么，才能称为 AI 交易 conviction？", "answer": "需要可证伪的 edge、机制、市场和时间尺度、反证、失效条件、基准、真实数据截止点、成本、容量、评估指标和风险边界；在这些被延迟验证前，它只是研究假设。", "source_ids": [ "damodaran-conviction", "nature-effective-learning" ] }, { "id": "retrieve-learning-20260813-application-explain", "kind": "application", "prompt": "如果要向普通人解释本次认知科学最重要的思想，你会怎么说？", "answer": "“别用刚读完时的熟悉感判断自己会不会。合上材料，隔一段时间先凭记忆说出来，再用反馈改正；真正学会还要能在一个没见过但相近的问题里用出来。”", "source_ids": [ "nature-effective-learning", "classroom-testing-meta", "transfer-review" ] }, { "id": "retrieve-learning-20260813-integrated", "kind": "application", "prompt": "本次三个学习标的分别应该用什么结构回忆？", "answer": "学习科学：机制—边界—测试；音乐史：年代—织体—机构网络；投资文章：定义—形成机制—行动后果—风险限定。", "source_ids": [ "nature-effective-learning", "met-renaissance", "damodaran-conviction" ] } ], "sources": [ { "id": "makeitstick-official", "title": "Make It Stick: The Science of Successful Learning", "url": "https://www.makeitstick.com/", "kind": "official book site", "evidence_role": "说明书籍的核心主张：检索、间隔、交错、反直觉的学习难度，以及重读造成的熟悉感问题。" }, { "id": "retrievalpractice-guide", "title": "Make It Stick — Retrieval Practice", "url": "https://www.retrievalpractice.org/make-it-stick", "kind": "authors’ research translation and teaching guide", "evidence_role": "补充书籍章节与检索、间隔、交错的实践说明。" }, { "id": "nature-effective-learning", "title": "The science of effective learning with spacing and retrieval practice", "url": "https://www.nature.com/articles/s44159-022-00089-1", "kind": "Nature Reviews Psychology review", "evidence_role": "学习科学综述，覆盖间隔、检索、元认知、保持和迁移。" }, { "id": "classroom-testing-meta", "title": "Testing (quizzing) boosts classroom learning: a systematic and meta-analytic review", "url": "https://pubmed.ncbi.nlm.nih.gov/33683913/", "kind": "systematic review and meta-analysis", "evidence_role": "课堂测试/测验增强学习的综合证据及其边界条件。" }, { "id": "interleaving-2024", "title": "Optimizing self-organized study orders: combining refutations and metacognitive prompts improves the use of interleaved practice", "url": "https://www.nature.com/articles/s41539-024-00245-7", "kind": "npj Science of Learning empirical study", "evidence_role": "说明交错与 blocked practice 的区别、学习者偏好及真实自组织学习中的条件性结果。" }, { "id": "desirable-difficulty", "title": "Getting the balance just right: Desirable difficulty in health professions education", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC10107184/", "kind": "open-access educational review", "evidence_role": "说明适宜困难的条件，以及先备知识不足时困难会变成负担。" }, { "id": "learning-definition-2026", "title": "Learning: One definition to rule them all?", "url": "https://www.nature.com/articles/s41539-026-00441-7", "kind": "npj Science of Learning cross-disciplinary perspective", "evidence_role": "截至本次记录较新的跨学科讨论，说明“学习”一词仍存在多种定义，不能把民间口号冒充统一科学定义。" }, { "id": "transfer-review", "title": "Mechanisms underlying training-induced cognitive change", "url": "https://www.nature.com/articles/s44159-021-00001-3", "kind": "Nature Reviews Psychology review", "evidence_role": "说明训练任务表现提升与迁移到未训练任务之间的争议，以及认知效率与一般能力的区分。" }, { "id": "prior-knowledge", "title": "Toward an understanding of when prior knowledge helps or hinders learning", "url": "https://www.nature.com/articles/s41539-021-00103-w", "kind": "npj Science of Learning research commentary", "evidence_role": "说明先备知识必须被激活、与任务相关并与新内容相容，才能稳定帮助学习。" }, { "id": "damodaran-conviction", "title": "The Situational Awareness Blow-up: The Collateral Damage from Investing Conviction!", "url": "https://aswathdamodaran.blogspot.com/2026/08/the-situational-awareness-blow-up.html", "kind": "Aswath Damodaran original investment commentary", "evidence_role": "核对 Leopold Aschenbrenner 的人名、Situational Awareness 论文名、investment conviction 的定义、来源和后果。" }, { "id": "oxford-plainchant", "title": "Western Plainchant", "url": "https://academic.oup.com/book/49864", "kind": "Oxford University Press music-history monograph", "evidence_role": "说明西方 plainchant 在加洛林文艺复兴后的 8—9 世纪开始书写，以及教会传统的长期演变。" }, { "id": "oxford-earliest-notations", "title": "The Oxford History of Western Music: Volume 1 — Music from the Earliest Notations to the Sixteenth Century", "url": "https://academic.oup.com/book/60817", "kind": "Oxford University Press music-history reference", "evidence_role": "说明 Gregorian chant 作为早期有文字记录的 repertory、早期复调和中世纪至文艺复兴的分期。" }, { "id": "cambridge-old-roman-chant", "title": "A new look at Old Roman chant", "url": "https://www.cambridge.org/core/journals/early-music-history/article/abs/new-look-at-old-roman-chant/4C57981EDCCFBE9E28A65AA8109EE394", "kind": "Cambridge University Press musicology article", "evidence_role": "区分 Old Roman 与 Gregorian chant 的传统、手稿年代和记谱证据。" }, { "id": "met-renaissance", "title": "Music in the Renaissance", "url": "https://www.metmuseum.org/essays/music-in-the-renaissance", "kind": "Metropolitan Museum of Art music-history essay", "evidence_role": "说明 1400—1600 的音乐、早期英国/北欧与勃艮第网络、法兰德斯作曲家、意大利交流和复调/音程发展。" }, { "id": "oxford-baroque", "title": "Baroque Music", "url": "https://academic.oup.com/reference/62400/reference-article-abstract/555404147", "kind": "Oxford Bibliographies in Music reference article", "evidence_role": "说明巴洛克时期的时间范围、复杂复调、装饰、体裁和功能调性的形成。" } ] }