How to Read a Classical Chinese Medicine Entry

How to Read a Classical Chinese Medicine Entry

The wall of characters has a skeleton

Open any entry in the Bencao Gangmu and the first impression is a solid wall of classical Chinese with no punctuation you recognise. But the wall has a fixed skeleton. Once you know the skeleton, an entry stops being a block of text and becomes a stack of labelled, attributed claims — each one traceable to who said it and when. This article reads one entry end to end so you can see the skeleton, and then read the next one yourself.

We will use licorice (gan cao) — one of the most-used herbs in the whole tradition, and a clean specimen because every part of the skeleton is present and legible. The point here is narrow: what are the parts of an entry, how do you tell the base classic from the later commentators layered on top of it, and can you check each quote against its source? It is a guide to reading the classics. It is not a buying guide, it carries no dosage, and it makes no claim about what gan cao does.

The name line: who logged it first, and at what grade

An entry opens with the herb's name, and immediately after it a small but load-bearing tag:

甘草《本經·上品》

That tag — 《本經·上品》 — is the first thing to read. It says gan cao's earliest record is in the Shennong Bencao Jing (the Ben Jing, the oldest of the base classics), and that the Ben Jing filed it in the upper grade (上品). Every entry carries a tag like this. It is, in effect, the citation for the herb's oldest layer — and it tells you which of the later text is commentary on top of that base, not the base itself.

释名 (names): even the aliases are attributed

The next section, 【释名】, lists the herb's other names, and here is the first lesson: the names are not anonymous folklore. Each is attributed. Gan cao's aliases — 蜜甘, 蜜草, 美草, 蕗草, 灵通, 国老 — each carry a source in parentheses (many marked 《别录》, the Mingyi Bielu). And when a name needs explaining, a named voice does the explaining. Tao Hongjing (陶弘景, 5th–6th century) glosses the striking alias 国老 ("state elder"):

國老即帝師之稱，雖非君而為君所宗。

"'State elder' is the title of an emperor's teacher — not itself the sovereign, yet honoured by the sovereign." You do not have to care about the metaphor. Notice the mechanism: the alias is Tao Hongjing's to explain, tagged as his, datable to his century — not a floating fact.

集解 (collected explanations): watch a later reader correct an earlier one

The 【集解】 ("collected explanations") section gathers what successive authorities said about where the herb grows, what it looks like, and how to grade it — stacked in rough chronological order, each clause tagged with its speaker (〔弘景曰〕, 〔苏颂曰〕, 〔李时珍曰〕…). This stacking is the whole reason source-traceability matters, because the authorities correct each other, and you can watch it happen.

Here is a worked example. Li Shizhen (李时珍, the Bencao Gangmu's own compiler, 16th century) quotes Shen Kuo's Mengxi Bitan to overturn an earlier commentator's identification:

按沈括《筆談》云：《本草》注，引《爾雅》蘦大苦之注，為甘草者，非矣。郭璞之注，乃黃藥也，其味極苦，故謂之大苦，非甘草也。

Unpacked: an old commentary on the classics had glossed the Erya's term 蘦 ("great bitter," 大苦) as gan cao. Shen Kuo (via Li Shizhen) says that is wrong — Guo Pu's note actually describes huang yao, which is intensely bitter, hence "great bitter," and gan cao (whose whole identity is sweet) cannot be it. This is the base-classic-versus-commentary distinction in one paragraph: a claim, its source, an earlier reader who got it wrong, and a later reader catching the error in the open — without deleting the original note. That is exactly what you are reading for.

气味 (nature and flavour): the line that also carries toxicity

Now the line most people come to an entry for — 〔气味〕, "nature and flavour." For gan cao it reads:

甘，平，無毒。

Three claims in four characters: the flavour is sweet (甘), the nature is neutral/balanced (平), and — crucially — the toxicity verdict rides on the same line: 无毒, "non-toxic." This is where to read the 毒 ("toxicity") judgement: it is not a separate warning box, it is the last clause of the nature-and-flavour line. When an entry instead reads 有毒 ("toxic") or 大毒 ("greatly toxic"), that word sits in exactly this slot. (Reading that single word correctly is load-bearing enough that MateriaTrace grades every herb's toxicity flag off precisely this clause.)

And even this base line attracts commentary. Kou Zongshi (寇宗奭, 12th century) immediately qualifies it:

生則微涼，味不佳；炙則溫。

Raw gan cao is "slightly cool and poor-tasting"; roasted, it turns "warm." So the flat 平 ("neutral") of the base line is already being argued with by a later voice who says it depends on preparation — and both are kept, side by side, each tagged.

归经 (channel attribution): a later layer, and the commentators disagree

Here is the subtlest lesson, and the one a beginner most often misreads. The channel attribution (归经) — which of the body's channels a herb is said to "enter" — is not in the base classics at all. The Ben Jing and the Bielu give flavour, nature, and indications, but no 归经. Channel attribution is a later development, largely a Jin–Yuan-dynasty (12th–14th century) contribution, and it enters the entry only through named commentators. Watch two of them attribute gan cao to different channels:

氣薄味厚，升而浮，陽也。入足太陰厥陰經。

That is Wang Haogu (王好古, Yuan dynasty): gan cao "enters the Foot Taiyin and Jueyin channels." But two clauses later, Li Shizhen gives a different reading:

通入手足十二經。

Li Shizhen: it "enters all twelve channels of the hands and feet." These do not agree — one says two specific channels, the other says all twelve — and the entry keeps both, each attributed. If you read only one and treat it as "the" channel of gan cao, you have flattened a genuine disagreement into a false certainty. The traceable way to read 归经 is: whose attribution, from which layer of the tradition — and to notice that the base classics did not make this claim at all.

主治 and 发明: the same attribution rule, all the way down

The two big remaining sections are 【主治】 (the indications — what the tradition recorded the herb as used for) and 【发明】 (the "elaborations," longer discursive commentary). We are not going to reproduce gan cao's indication list here — this is a guide to reading the entry, not a therapeutic text, and MateriaTrace makes no claim about what any herb does. But the structural point is the same as everywhere else: in 【主治】 every clause is tagged to its source, and you can see the layers — the oldest indications marked 《本经》, then 《别录》, then named physicians (甄权, 大明, 李杲, 好古, 时珍). The rule never changes: each claim to its own source. That is what lets a later reader — or a database — keep the Ben Jing's layer distinct from a 16th-century addition instead of blending them into one undated paste.

One more thing lives in this stretch that is pure reading-skill, not usage: the "seven relations" note. Xu Zhicai records which herbs gan cao is helped by, dislikes, or is opposed to:

術、苦參、乾漆為之使，惡遠志，反大戟、芫花、甘遂、海藻。

The four herbs after 反 ("opposes") are the classic incompatibility record — gan cao "opposes" da ji, yuan hua, gan sui, and hai zao. That is a textual relationship between herbs, logged in the entry; we treat it as exactly that, a historical record, and we wrote a separate article on those incompatibility lists. It is not advice to the reader about combining anything.

How MateriaTrace holds it: the skeleton, kept intact

So reading a classical entry is really reading a labelled skeleton:

1. name + grade tag (甘草《本經·上品》) — which base classic first logged it, at what grade;
2. 释名 — the aliases, each attributed;
3. 集解 — provenance and grading, stacked by author, where you watch later readers correct earlier ones;
4. 气味 — flavour, nature, and the toxicity verdict, all on one line;
5. 归经 — a later layer, present only through commentators, who may disagree;
6. 主治 / 发明 — indications and elaboration, every clause tagged to its source.

Here is how we hold that on the page: each line is kept verbatim, attributed to the authority who said it (弘景, 沈括, 王好古, 时珍…), placed in its section of the Bencao Gangmu, and carries a content fingerprint so you can open the original book and confirm the text is there, exactly, at that spot. Where the authorities disagree — neutral versus preparation-dependent, two channels versus twelve — we keep the disagreement visible rather than picking a winner.

Worth stating plainly: this article explains how to read the structure of a classical materia medica entry, for study and traceability. It is not advice on using, preparing, or dosing any herb, and it makes no claim about gan cao's effects. MateriaTrace records what sources say and makes no treatment recommendation; any question about a herb belongs with a qualified professional.

Why reading it this way is the whole point

You could take an entry and boil it down to one modern-sounding line — "licorice: sweet, neutral, non-toxic, enters all twelve channels" — and it would read cleanly. But that line throws away everything the entry actually is: a base-classic tag you can cite, a correction Shen Kuo made to Guo Pu that you can trace, and two commentators openly disagreeing about the channels. Those survive only if the exact lines survive, each attributed, each checkable — so that you can do what Li Shizhen did, and read the record against itself.

See it on the page: open licorice (gan cao) and watch each field trace back to its named source, and huang yao zi, the herb Shen Kuo said had been mistaken for it. For a single herb read as a verification problem, see how the classics verify real ginseng; for what happens when the sources themselves disagree on the toxicity line, read about cinnabar and mercury.

If this "every line clicks back to its own source" approach is useful to you, see what else is in MateriaTrace on the homepage.

MateriaTrace is a source-traceable, bilingual knowledge graph of Chinese medicine. It records how classical and modern sources describe each entry — each field traceable to its primary source, contradictions kept side by side — and makes no treatment or usage recommendations. Every line quoted in this article can be traced to its classical source.