我们怎么核对一篇稿子里的每一句话

▍一、三级来源，只认头一级

所谓三级来源，指的是按可信度把材料分为官方原文、二次报道、聚合内容三档。头一档是官方原文：政府网、主管部门网站、名录数据库、官媒通稿。第二档是二次报道与转载。第三档是自媒体与聚合内容。

写稿时只用头一档。第二档的用途只有一个：当线索，顺着它回头找原文。第三档不进正文，看都不作为依据。

▍二、数字必须带日期和出处

同一项数据在不同年份、不同部门的口径里常常对不上。我们的处理办法不是挑一个看起来顺眼的，而是把日期和来源一起写进句子里：据某部门某年某月的公开报道。

如果两条权威口径互相冲突，就把两条并列写出来，或者干脆不写这个数字。宁可少一个数据，也不让读者拿走一个含糊的印象。我们常用的标注格式是「据某部门某年某月的公开报道」；例如，据山西省文化和旅游厅 2026 年 10 月 27 日报道，新修订的《山西省非物质文化遗产条例》已于 2026 年 6 月 1 日施行——这类句子带着日期与发布主体，读者可以自己去核。

摊开的笔记与打印资料，页面上划着待核实的段落

▍三、答不上来的，就写「据公开资料」

有一类信息很难查实：某个工坊的具体人数、某年的确切产值、某项技艺在某地的流传年限。这类内容我们的规矩是，要么找到一手，要么写明「据公开资料」并注明出处。

我们内部还有一份「待核实清单」，凡是没有双源互证的数字，一律进清单，不出稿。清单里的条目会一直挂着，直到某天被证实或者被放弃。

▍四、品牌的话，只说已经确认的

最难核对的其实是我们自己。门店数量、合作的匠人人数、海外布局的规模，这些数字外界常替我们添。我们的做法是：没有确认过的，一律不写；能写的，也不加码。

行业里对工作边界早有定论，写在法里：

国家对非物质文化遗产采取认定、记录、建档等措施予以保存。

（引自《中华人民共和国非物质文化遗产法》第三条）

这不是谦虚，是算账。我们在山西平遥的漆坊里问过大漆的髹涂层数，得到的答案每次都要再核一遍。一个靠信任做买卖的品牌，说错一句话的成本，比少说一句高得多。

▍问答

问 ｜ 做非遗内容时，怎么判断一个来源可不可靠

答 ｜ 看三件事：能不能追到官方原文、有没有标注日期与发布主体、数字是否能与另一份权威材料互证。转载与自媒体只能当线索，不能当依据；同一事实出现两种官方口径时，并列写出或干脆不写具体数字。

问 ｜ 为什么很多非遗文章里的数字对不上

答 ｜ 常见原因有三个：统计年份不同、统计口径不同（是否包含省市级项目、是否含工坊与从业者）、发布主体不同。所以引用时必须带日期与来源；看到没有出处的数字，先当作未核实处理。

问 ｜ 韵和鸣为什么会长期做匠人访谈

答 ｜ 因为手艺的细节只在匠人嘴里。工序为什么这样安排、材料为什么这样取舍，不坐下来问就写不出来。我们有自己的匠人访谈栏目，帮老师拍产品、拍视频，把手艺的过程记录下来，再用多语言直播与翻译讲给国内外的读者听。

📩 非遗项目 ｜ 匠人举荐 ｜ 跨境出海 ｜ 线下联营

欢迎留言或后台联系，我们与匠人同行。

—— 韵和鸣 · 与匠人同行

〖 生 〗

English

Three Tiers of Source, and Only the First Counts

By three tiers of source we mean sorting material by reliability into official text, secondary reporting and aggregated content. The first is official text: government portals, department websites, listing databases, agency press releases. The second is secondary reporting and republication. The third is self-media and aggregated content.

Only the first tier is used in drafting. The second has one use: as a lead, followed back to the original. The third never enters the text and is not treated as evidence.

The same figure often disagrees between years and between departments. Our handling is not to pick whichever looks tidy, but to write the date and source into the sentence: according to a public report by such-and-such body in such-and-such month.

If two authoritative sources conflict, we set both side by side, or drop the figure altogether. Better one datum fewer than leaving a reader with a vague impression. Our usual format runs "according to a public report by such-and-such body in such-and-such month"; for instance, the Shanxi Provincial Department of Culture and Tourism reported on 27 October 2026 that the revised Regulations of Shanxi Province on Intangible Cultural Heritage took effect on 1 June 2026. A sentence carrying a date and an issuing body is one a reader can check independently.

Open notebook and printed sources, paragraphs marked for verification

When We Cannot Verify, We Write "According to Public Sources"

Some information resists verification: the exact headcount of a workshop, a precise output value for a given year, how long a technique has circulated in a place. Our rule for these is either find a primary source or write "according to public sources" and name where it came from.

We also keep an internal list of items pending verification: any figure without two corroborating sources goes on it and stays out of drafts. Entries hang there until they are confirmed one day, or abandoned.

About Ourselves, Only What Has Been Confirmed

The hardest thing to verify is ourselves. Store counts, the number of makers we work with, the scale of our overseas presence: others often add to these for us. Our rule is simple: nothing unconfirmed gets written, and what is written is not inflated.

E> The state preserves intangible cultural heritage by means of recognition, documentation and archiving.

The field settled its working boundaries long ago and wrote them into law: recognition, documentation and archiving are how the state preserves heritage. This is not modesty, it is arithmetic. In the lacquer shops of Pingyao, Shanxi, we have asked how many coats of raw lacquer go onto a piece, and every answer had to be checked again. For a brand that trades on trust, the cost of one wrong sentence far exceeds that of one sentence fewer.

How do you judge whether a source on heritage is reliable?

Three things: whether it traces back to an official text, whether it carries a date and issuing body, and whether the figures are corroborated by another authoritative document. Republication and self-media are leads, not evidence; where two official versions of the same fact exist, set both out or omit the number.

Why do figures in heritage articles so often disagree?

Three common reasons: different years, different definitions of what is counted, and different issuing bodies. Figures must therefore carry a date and source; a number without a source should be treated as unverified.

Why does YUNHEMING keep running interviews with makers?

Because the details of a craft live in the maker's account. Why a step is ordered thus, why a material was chosen over another, cannot be written without sitting down and asking. We run our own interview series, filming products and processes for the masters, and then carry the craft to readers at home and abroad through multilingual livestreams and translation.

ICH Projects · Artisan Referrals · Global Distribution · Salon Partnerships