HN 写作课

工程师们二十年里聊过的写作。

工程师、作家、编辑和创业者在 Hacker News 上聊了二十年“怎么写”:怎么写得清楚、怎么讲好一个故事、技术文档怎么写、小说和出版是怎么回事。这里把其中关于写作与讲故事的帖子挑出来,配上中文导读、可以直接用的写作要点和评论翻译。

从必读开始
  • 2,600+篇关于写作的 HN 讨论

Writing tips

写作锦囊

正在抽一条写作要点…

Essential

必读 30 篇

在讨论库里看全部精选

编辑精选里 HN 讨论最热的 30 篇,每篇都有中文导读和写作要点。

  1. 正在载入…

Archive

讨论库

按发帖年份分布,点柱子筛选。

排序

    About

    这些讨论是怎么挑出来的

    1. 01

      从全量里捞

      读取 Hacker News 自 2006 年以来的全部帖子,用关键词先捞出 40 分以上、可能和写作有关的约 1.1 万篇。

    2. 02

      只留真正讲写作的

      由 AI 逐条判断是否和写作、讲故事、文风、出版有实质关系,剔除“写代码”“编辑器”这类字面撞车,并分进十个类别。

    3. 03

      提炼可用的要点

      为值得读的帖子写中文导读,提炼 2-4 条可以直接用到自己写作里的要点,并精选翻译 HN 高赞评论。内容由 AI 生成,可能有误。

    语言与文风2016theatlantic.com

    道本语(Toki Pona):一门只有100个词的人造语言

    Toki Pona: A created language that has only 100 words

    阅读原文 HN 讨论
    132HN 得分
    88条评论
    4条写作要点

    导读

    这篇文章介绍 Toki Pona:一种只有约 100 个词的人造语言,其核心理念是“少即是多”,用极简词汇加组合来表达复杂意思。对写作者来说,它是一次关于“限制如何塑造表达”的思想实验:词汇越少,越考验你如何把意思拆开、重新组合,也越容易暴露歧义。

    关键词 托基波纳人造语言简洁表达语义

    可以直接用的写作要点

    1. 01

      把复杂意思拆成少数基本概念再组合表达,往往比堆砌大词更清楚。

    2. 02

      少用一词多义、依赖语境的词(如 cool、blue),它们看似省事,实际会让读者猜测。

    3. 03

      给文章里的关键概念下定义时,避免用被定义的词解释自己,否则容易循环定义。

    4. 04

      给自己设词汇或篇幅限制,是逼出简洁表达的有效练习。

    HN 评论精选

    一个词承载多重含义会牺牲精确性,所谓的“简单”反而制造理解负担。

    关键在简单。而在 Toki Pona 里,“简单”字面上就等于“好”,两个概念合并成一个词 pona。英语里也有类似情况:cool 指温度低,但用来形容人时又不是;blue 指一种颜色,但涉及情绪时又不是。这绝不简单,老实说我也不觉得好听。

    yongjik查看英文原文

    > The point is simplicity. And in Toki Pona, simple is literally good. Both concepts are combined in a single word: pona. That's how you end up with words like English "cool", which means low(-ish) temperature, except when it's about people, or "blue", a particular color, except when it's about feelings. It's anything but simple, and honestly it doesn't sound good to me, either.

    词义可以拆成少量语义维度,核心符号加组合规则就能覆盖全部表达;辅助语言应优先考虑易学。

    这让我想到自然语言处理工具 word2vec:它把词变成向量,每一维代表词义的某个方面(比如是否指地点、是男是女),这些向量不是人设计的,而是从数据里学出来的——出现在相似语境里的词,向量也相似。于是可以做出“国王-男人+女人=女王”这样的运算,甚至能帮机器翻译:把两种语言里几个对应的词固定成同一向量,其余词也会靠拢。说到底,具体的词和符号并不重要,重要的是准确表示词义所需的那约 100 个维度;只是人说出一个百维向量,远比说一个词低效。但借此也许能找出表达所有词所需的最小符号集和组合规则。很多人抱怨这不像自然语言,可辅助语言的重点本就是好学易用。

    Houshalter查看英文原文

    This reminds me a lot of a natural language processing tool called word2vec. It reduces words into vectors, where each dimension represents something about the meaning about the word. E.g. whether it's a place, or if it's male or female. And these vectors are not designed by humans, but learned from data. It's optimized so that words that occur in similar contexts have similar vectors. E.g. places tend to occur near words talking about places, like "there" or "go", and male names tend to occur near "he" and "his", etc. You can do really interesting things with these vectors. Like 'King'-'man'+'woman'='queen'. Recently a computer program beat the word analogy part of an IQ test using this. It's useful in computer translation. Take two languages and constrain a few identical words to have the same vector. Then all the other words also end up with similar vectors. Anyway, with something like this, the actual words and symbols don't matter, just the 100 or so dimensions needed to accurately represent the meaning of every word. Of course speaking a 100 dimensional vector is way less efficient than a single word or symbol. But perhaps using this, we could come up with an absolute minimal set of symbols necessary. And some rules for combining them to get all words. I know a lot of the comments here are complaining that this doesn't make a good natural language. But that's not the point of an auxiliary language. The main thing auxiliary languages should optimize for is how easy it is to learn. Simpler is better. Reducing the words and grammar is hugely important. Another advantage of such an approach, is that it might be more amenable to machine translation. Which is another advantage for an auxiliary language.

    用有限“语义基元”定义所有词汇,可以避免循环定义,并让隐藏的歧义显现出来。

    类似的项目还有语言学家在 70 年代初提出的“自然语义元语言”,它由约 70 个“基元”组成,其他所有词都能分解成这些基元——就像写词典时必须避免用被定义的词来解释自己。与 Toki Pona 等语言不同,它主要用于思想实验,以及分析现有语言中的歧义和双重含义。

    zachrose查看英文原文

    Cool. There's a similar project put forth by linguists in the early 70s called Natural Semantic Metalanguage, with about 70 "primes" that all other words can be factored out as, like you would have to do if you wrote a dictionary without circular definitions. Unlike Toki Pona or the other languages mentioned in the article, Natural Semantic Metalanguage is mostly used for thought experiments and to analyze existing languages for ambiguities and double meanings.* https://en.m.wikipedia.org/wiki/Natural_semantic_metalanguage *"Double meanings" may not be the right word, but is there a word to describe what happens when a system is non-orthagonal and the same piece winds up with multiple responsibilities?

    同类还值得读

    1. 像说话一样写作

      Write like you talk

      这篇短文主张写作应像说话一样自然、清晰,避免装腔作势的书面腔。

      语言与文风2015 · 647 分 · 272 评4 条写作要点编辑精选
    2. 法律语言难懂源于写作不佳,而非专业概念

      Poor writing, not specialized concepts, drives difficulty with legal language

      这是一项认知科学研究:法律文本难懂,主要不是因为充斥专业术语,而是因为写作者把句子写得又长又绕、修饰语层层嵌套、回指距离过远。

      语言与文风2022 · 559 分 · 322 评5 条写作要点编辑精选
    3. 写得简单

      Write Simply

      保罗·格拉汉姆在这篇短文里主张:好文章应该写得简单——用日常词汇、短句、直白的结构,不为显得聪明而堆砌复杂表达。

      语言与文风2021 · 484 分 · 330 评4 条写作要点编辑精选
    4. Tortured phrases:科学写作中出现的一种可疑文风

      Tortured phrases: A dubious writing style emerging in science

      《自然》新闻报道了一种在科学论文中蔓延的怪现象:原本标准的术语被同义替换软件改写成了古怪说法,比如把 "cloud computing" 写成 "haze figuring"。

      语言与文风2021 · 329 分 · 256 评4 条写作要点编辑精选
    5. 乔治·奥威尔 1940 年对《我的奋斗》的书评

      George Orwell's review of Mein Kampf (1940)

      这是奥威尔1940年对《我的奋斗》的书评。

      语言与文风2014 · 305 分 · 240 评4 条写作要点编辑精选

    打赏支持

    打赏二维码

    觉得有用,请作者喝杯咖啡。

    WeChat

    关注向阳乔木

    微信公众号「向阳乔木推荐看」

    微信公众号「向阳乔木推荐看」二维码

    微信扫一扫;在微信里打开本页时,长按二维码识别。