记忆宫殿 地点法 History 时间线 地理案例
考前一晚翻开 History 笔记，一页上是 1867、1871、1878 三个年份，另一页是五个条约的名字。再翻到 Geography，是里约、拉各斯、布里斯托三个案例，加上它们的人口数和降雨量。很多学生以为这类内容只能靠反复抄写解决，抄到第三遍手酸了，第二天做真题还是想不起来。地点法（method of loci，中文习惯叫记忆宫殿）处理的正是这种一串彼此不相干的信息。把每个条目放到一个你已经走熟的空间位置上，回忆时按路线走一遍，条目跟着位置一起回来。它不替代理解，也不替代做题，它解决的是记住了却在考场上取不出来这一段。
Open a History folder the night before a mock and the page holds three treaty dates and five names. Open the Geography one and it holds Rio, Lagos and Bristol with their population figures. Students often assume the only way through material like this is to copy it out again until the hand aches. The method of loci, usually called a memory palace in Chinese classrooms, deals with exactly that kind of list: unrelated items pinned to positions along a route you already know, which come back when you walk it. It does not replace understanding, and it does not replace past papers.
地点法从古希腊的宴会厅开始
地点法有一个流传很广的故事。公元前 5 世纪，诗人西摩尼德斯（Simonides of Ceos）在一次宴会屋顶坍塌后，靠记住每位客人坐的位置认出了遇难者。这个故事更像传说，真正把它写成技术的是古罗马的修辞学手册《Rhetorica ad Herennium》，成书于公元前 80 年代左右。那本手册教演讲者把论点的要点依次放进一栋建筑的房间，讲的时候沿着房间走。所以地点法最早不是为考试设计的，而是为了不打草稿讲一段长话。今天用它背时间线和案例，骨架没有变。
The technique is usually traced to a story about the poet Simonides of Ceos, who could name the dead after a roof collapsed on a banquet because he remembered where each guest had been sitting. What turned it into a working method was a Latin handbook on rhetoric, Rhetorica ad Herennium, written around the 80s BC, which told a speaker to place the points of a speech in the rooms of a building and walk through them while talking. It was built for speaking without notes rather than for exams, and the skeleton has not changed since.
熟悉的空间为什么比年份好记
大脑对位置的记忆和对年份的记忆不是一套系统。1971 年，O'Keefe 与 Dostrovsky 在大鼠海马体里记录到随位置变化的放电，这类细胞后来被称为位置细胞（place cells）。这项工作与后续研究让 O'Keefe 和 Moser 夫妇获得 2014 年诺贝尔生理学或医学奖。位置信息在大脑里有专门的硬件，年份和条约名称没有。把 1871 挂到楼梯转角，等于借了这套硬件的力。很多学生以为记不住是因为背得不够用力，其实是编码通道选错了。
Place memory and date memory do not run on the same machinery. In 1971 O'Keefe and Dostrovsky recorded cells in the rat hippocampus that fired as the animal moved through particular spots, later named place cells, and that line of work contributed to the 2014 Nobel Prize in Physiology or Medicine. Location has dedicated hardware behind it, while a treaty date does not, so hanging 1871 on a stair turn borrows hardware that would otherwise sit idle.
六周训练能改变的不只是分数
地点法不只有轶事支撑。2017 年，Dresler 等人在《Neuron》发表了一项训练研究。参加者包括世界级记忆运动员和未训练的对照组，普通参与者做了约六周、每天约 30 分钟的地点法训练。训练后他们的回忆成绩接近运动员水平，静息态脑网络连接也向运动员的模式靠拢。样本不大，结论不宜夸大，但它说明地点法是可以练出来的技能。对备考的含义很直接：前三周不进考场不用慌，这套动作需要时间才成型。
The method has more than anecdote behind it. In 2017 Dresler and colleagues published a training study in Neuron involving memory athletes and untrained controls, and the novices trained for roughly six weeks at about thirty minutes a day. Their recall rose close to athlete level, and resting-state connectivity shifted toward the athletes' pattern. The sample was small, so the result should not be oversold, yet it does show that the skill can be trained rather than being innate.
一条走廊装一个世纪的时间线
History 的时间线是一维的，宫殿是三维的，这个差别正好用来装结构。把一栋你自己走熟的楼分成段落：一楼放 19 世纪后半段，二楼放 1900 到 1945，走廊尽头放冷战。每层楼的房间数要和这一段里的关键事件数对上，不要硬塞。走到楼梯口就是一个转折点，回走时你会先遇到它。很多学生以为时间线要按年份等距排列，其实按因果和转折排更牢。排完之后你的笔记顺序就是卷面上可能出现的顺序。
A timeline is one-dimensional and a building is not, and that mismatch is useful. Divide a building you already know into periods: the ground floor for the late nineteenth century, the first floor for 1900 to 1945, the end of the corridor for the Cold War. Keep the number of rooms matched to the number of key events in each stretch instead of cramming. A stair turn then becomes a turning point, and you meet it first on the way back.
条约和政策要挂在一件具体物件上
年份好放，人名和条款难放，因为它们没有形状。做法是给每条加一件能看见的物件。1871 年的《法兰克福条约》可以放在一张被割走一角的桌子上，某条改革法案可以放在一只被掰断的笔上。物件不必合理，越奇怪反而越容易回来。评分方案看的是史实是否具体、人名和年份是否准确，图像只是帮你把这两样取出来。常见的错误是把整段笔记塞进一个房间，房间太挤，什么也拎不出来。一个房间放三到四条就够。
Dates fit easily; names and clauses do not, because they have no shape. Attach an object to each one: the 1871 Treaty of Frankfurt on a table with a corner sawn off, a reform act on a snapped pen. The object does not need to make sense, and the stranger it is the more reliably it returns. Mark schemes reward specific and accurate detail, and the image exists only to retrieve that detail. Stuffing a page of notes into one room is the usual mistake, because the room gets crowded and nothing can be lifted out.
地理案例一个房间放一个地方
Geography 的案例研究是地点法最省力的用法，因为案例本身就是地点。AQA GCSE Geography 8035 的考纲要求答案里出现具名的地点，不能只写某个发展中国家。把里约放厨房，把拉各斯放浴室，把布里斯托放阳台。每个房间固定放三样：这个地方的特征、一个能背出的数字、一条评价，比如城市更新让谁受益、谁被迁走。走一遍房间，三样东西按顺序出来。答题时就不会只剩一个地名和一句空泛的总结。
Case studies are where the technique pays best in Geography, because a case study is already a place. The AQA GCSE Geography 8035 specification expects named locations in answers rather than a generic reference to a developing country. Rio goes in the kitchen, Lagos in the bathroom, Bristol on the balcony, and each room holds three things: the character of the place, one number you can quote, and one evaluation such as who gained from regeneration and who was moved out.
数据先压成一个图像钩子
数字最难留，因为它们彼此太像。1400 万、1.4 万、41% 放在一起，看三遍也分不清。做法是先把数字变得夸张。人口 1400 万可以变成一个挤满人的地铁站台，41% 可以变成四根柱子断了一根。概念也能这样处理，需求价格弹性大的商品可以是一个被捏扁的易拉罐。写答案时再把图像还原成数字，注意单位。这一步最容易出错的是数字还原，写完回头核一遍，不要把人每平方公里写成万人。
Numbers are the hardest part to hold, because they look alike. Fourteen million, fourteen thousand and forty-one per cent blur together after three readings. Exaggerate first: fourteen million becomes a platform packed with people, forty-one per cent becomes four pillars with one snapped. Concepts work the same way, and a price-elastic good can be a can crushed flat. Then reverse the image into a figure when you write, and check the unit, because persons per square kilometre turning into ten thousand people is the slip that costs marks.
按试卷顺序排宫殿的动线
宫殿的路线最好和卷面顺序一致，这样打开试卷就知道从哪走。AQA 8035 是三张卷。Paper 1 考自然环境，1 小时 30 分钟、88 分、占 35%。Paper 2 考人文环境，1 小时 30 分钟、88 分、占 35%。Paper 3 考地理应用与野外调查，1 小时 15 分钟、76 分、占 30%。Edexcel A-Level History（9HI0）是四部分：Paper 1 占 30%，Paper 2 占 20%，Paper 3 占 30%，课程作业占 20%。按这三卷或四部分分段，比按教材章节分段更贴近考试。
Match the route to paper order where you can, so that opening the paper tells you where to start walking. The AQA 8035 Geography suite has three papers: physical environment at one hour thirty and 88 marks for 35 per cent, human environment at the same length and weight, and geographical applications at one hour fifteen and 76 marks for 30 per cent. Edexcel A-Level History, 9HI0, is four components, with Paper 1 at 30 per cent, Paper 2 at 20, Paper 3 at 30 and coursework at 20. Sectioning by paper beats sectioning by chapter.
每次重走要卡住间隔
建好宫殿不等于记住，重走的间隔比次数更影响结果。2008 年，Cepeda 等人在《Psychological Science》发表的研究给出一个可用的经验值：复习间隔取目标记忆期限的 10% 到 20% 比较合适。如果考试在三个月后，两三个星期重走一次就够；如果只剩一周，每天走一遍。很多学生以为走得越勤越好，结果把力气都花在考前几天。重走时不要看笔记，先在脑子里走完再核对，漏掉的房间单独修。间隔对了，同一个宫殿能撑整个学期。
Building the palace is not the same as holding it, and spacing matters more than repetition. A 2008 study by Cepeda and colleagues in Psychological Science suggested a workable rule: set the gap between reviews at roughly ten to twenty per cent of how long you need to remember the material. Three months out, walking it every two or three weeks is enough; one week out, walk it daily. Review without notes first, then check what went missing.
宫殿别建太大
地点法最容易失败的地方是规模。有人一上来就想象一座城堡，二十个房间，每个房间八个格子，第二周就再也走不下来。能稳定复述的路线大概十个站点，超过这个数就要靠运气。用你已经走过上百次的地方：自己的家、教室到食堂的路、亲戚家的院子。新建的虚构宫殿需要反复巩固，成本高于收益。宁可建两个小宫殿，一个管时间线，一个管案例，也不要建一个大而全的。建完第一次就走一遍，走不下来的当场拆掉重来。
Scale is where the technique usually breaks. A castle with twenty rooms and eight slots each looks impressive and becomes unwalkable within a fortnight. A route you can recall reliably holds around ten stops, and beyond that you are guessing. Use somewhere you have walked hundreds of times, such as your own flat, the route from classroom to canteen, or a relative's yard. Two small palaces, one for the timeline and one for case studies, work better than one large one.
图像不具体是最常见的失败
最常见的错误是图像不够具体。写一张桌子没有用，要写一张桌面裂开、堆着旧报纸的桌子。另一个毛病是走过场，脑子里的确走了一遍，但每个房间只看一眼，细节全空。判断会没会有一个简单办法。合上笔记，从门口走到最后一个房间，把每个房间的三样东西说出来，说不出话的那个房间就是空编码。补的时候不要重走全部，只修那两个房间。这一步最容易省，也最容易在考场上暴露。
Most failures come from images that are not concrete enough, or from a walk-through performed too quickly. A table is not an image; a table with a split top and stacked newspapers is. Walking the route while barely looking at each room leaves hollow slots, and the hollowness only shows up in the exam. Close the notes, walk from the door to the last room, and say the three things in each; whatever room goes silent is the one to rebuild.
解释题和因果链不适合放宫殿
地点法擅长取回事实，不擅长生成论证。地理的解释为什么题、历史的史料解读题，考的是因果链和判断，把因果拆成图像反而更慢。AQA 与 Edexcel 的解释题都按论证层级给分，需要的是把原因串成逻辑，不是把名词摆出来。所以宫殿只负责前半程，把该有的事实备好，剩下的交给练习测试。2013 年，Dunlosky 等人在《Psychological Science in the Public Interest》的综述里评了十种学习法，练习测试和分散练习被评为高效用，重读与划重点属于低效用。宫殿属于前者的辅助工具。
The method retrieves facts; it does not build arguments. An explain why question, or an interpretations question on a source paper, is graded on the chain of reasoning, and turning that chain into pictures is slower than writing it out. Both AQA and Edexcel mark those answers by level of argument rather than by the number of names on the page. Let the palace carry the facts and hand the reasoning to practice testing; Dunlosky and colleagues, reviewing ten techniques in 2013, rated practice testing and distributed practice as high utility and rereading as low.
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