Week 12 Session: The taxonomic congress

Entries are tabled and challenged on one ground only, an undeclared field. 猼訑 shows what the rule does to a shared trait, 窫窳 forces a pick nothing justifies, and the course answers its own question.

24 May 2027 · Phase 3, Take it apart and rebuild

What this week does#

Week 11 asked which of your characters survive contact with material the edition did not write. This is the last week, and it does two things: it admits entries under one rule, and it settles the account for the whole system — every field the course built, and whether that field held.

The rule has two clauses and the second one is the whole course. Citing a line is easy and everyone manages it. Declaring which of your own fields that line does not supply means writing down, inside your own submission, the list of places where you decided rather than read.

Five sections. Three worked cases, then the account, then the rule:

1. A trait shared by three unrelated creatures — 猼訑, and what the rule does to nine tails.
2. Two records that will not resolve — 窫窳, where you must pick a body you cannot justify.
3. An entry form in the wrong order — 蛊雕, whose record arrives back to front.
4. The account, which carries no creature of its own because it is about the twelve weeks rather than about any one record.
5. The rule the congress runs on, with 顒 as the last test of it, and the course’s answer to its own question.

Nothing else is grounds for rejection today. An entry may be eccentric, badly-argued or wrong. It is admissible if it is honest about its gaps.

Example 1: nine tails on three unrelated creatures#

猼訑 (bóshī) is a sheep with nine tails, four ears and eyes in its back, at 基山. 陆吾 has nine tails on a tiger’s body at 昆侖之丘. The unnamed beast at 青丘之山 has nine tails on a fox. Three creatures, three chapters, nothing else in common, and no line connecting any two of them.

An entry treating nine tails as a taxonomic character is making a claim no record makes. It is still admissible — but only if it says so, in the declarations, in one sentence. An entry that quietly groups the three has smuggled a conclusion into a data field, where it stops looking like a claim and starts looking like an observation. The congress rejects that one.

Example 2: two records that will not resolve#

窫窳 (yàyǔ) is on the table deliberately. 海內南經 gives it a dragon’s head and a 貙’s body, in the 弱水, eating people. 海內西經 gives it a snake’s body, a human face, and a death. Week 8 established that these cannot be one animal.

So an entry on 窫窳 has to pick a body and then declare that the pick is unjustified, because nothing in either record ranks them. That is an uncomfortable sentence to write about your own work. Writing it is the assessment.

Example 3: an entry form in the wrong order#

蛊雕 (gǔdiāo) is named before it is described. In its record 名曰蠱雕, it is called 蠱雕, arrives ahead of 其狀如雕而有角, shaped like an eagle but with horns, reversing the order that every entry form in this course assumes — week 2 built its columns by ruling records into them left to right, and this record does not run left to right.

An entry using it has to declare that it reordered the source. It also has to decide whether 其音如嬰兒之音, its cry is like an infant’s, which it shares almost word for word with the 青丘 beast, is a character or a formula, and declare which way it went. Week 5 answered that question for omens. Nobody has answered it for voices.

The account: every field, and whether it held#

The entry form was not designed in one go. Each field arrived in a particular week for a stated reason, and most of them were tested to destruction in a later one. Here is the whole set, with the week that put it in and the week that settled it.

Read the settled column and the shape of the semester is in it. Everything that arrived in weeks 1 to 4 and survived is either a count or a citation. Everything withdrawn was withdrawn in the stress phase, and each was withdrawn for its own reason rather than for a general loss of nerve: the omen because it repeats, the use because it is about the reader, the kind because it is not a rank, the name because it does not pick out a thing.

Two rows are worth more than the rest. Citation held, which is why it is the one thing the congress requires. Name failed, which is the row that should worry anybody designing a classification of anything, because a name field is the one field every system starts with and the one nobody thinks to test.

The rule the congress runs on#

顒 (yú) is the last test. 見則天下大旱 — where it appears, the world goes dry — sits on it word for word and on five other records that share no part with it. Any entry filling an omen field is filling a field week 5 showed belongs to the formula rather than to the animal, and the declaration has to say so. An entry that fills it silently has asserted that the omen is a property of the bird, which is the one thing this course has established that it is not.

So: can a rigorous classification system be built for the creatures of the 山海经 (Shānhǎijīng, the Classic of Mountains and Seas)?

Partly, and the table above is the whole of the answer. Two fields held outright, one held and turned out to mean something else, one split in two, four were withdrawn or failed, and one is an admitted invention. Four usable fields out of nine is not a success. It is not a failure either. It is a result, and the difference is that every field that went is on the table with the record that took it down.

That is what the rule’s second clause is for. A taxonomy that hides which parts it made up is the thing this course is against, and a declarations section is how an entry stops doing it. The successes are the fields that held. The failures are the fields that did not, and they are the part worth submitting.

What you leave with#

A bestiary entry with a declarations section, and two reviews of other people’s. A review is marked on whether it found an undeclared field, not on its tone, and a review that finds none and explains where it looked is a full-marks review.

Then Part 2 of the classification design, which is the last piece due and the only one that leaves this book. Take the system you proposed in week 7, amended by whatever weeks 8 to 12 did to it, and run it on records this course never supplied. The congress is where the objections to it come from, which is why Part 2 is written after the congress and not before.

The transferable part is not the system. It is the habit of writing the declarations at the same time as the fields, so the list of places where you decided rather than read exists before anybody asks for it. That habit costs one sentence per field, and it is the only thing in this course that works on material nobody has classified yet.