Surface vs Deep Structure

Deep structure is the functional parse of a character, the parts doing sound or meaning work in this host. Surface structure is how the same graph is written and filed today: strokes, radicals, leftover marks, and the official standards. The cut stops at the parts that do a job here; the radical is a filing address, and neither layer is generative grammar.

The stopping rule

部 splits into 咅, sound only here, and 邑 compressed to 阝, meaning. 立 and 口 inside 咅 do nothing in 部. Cutting that deep yields noise.

A functional component is a part doing a job in this host. Four roles cover the jobs: form, meaning, sound, empty. Form plus meaning is the semantic pair. The first three are field-standard; empty is a functional verdict — no sound, no meaning. Non-functional is a different word: a sub-component below the stopping line, the way 立 and 口 sit inside 咅 in 部. Form Components owns depicting. Meaning Components owns sense. Sound Components owns pronunciation, including drifted phonetics. Empty Components owns the residue verdict. Three Attributes of a Character is the triad those roles express.

立 keeps its shape and changes its job. It is meaning in 端. It is meaning and sound in 位. It is sound in 粒 and 拉. It is empty in 音. It is non-functional in 部 and 章 — below the stopping line, not a residue verdict. Six hosts, one shape.

Component sense can differ from the standalone word. 而 as a component is a beard; in 端 the conjunction “and” plays no part. 而 and 山 are inert sub-components of 耑, the sound carrier in 端.

Opaque sound links still count. 立 in 位, 丁 in 成, 疋 (a variant of 正) in 定 are still doing sound work, even when modern ears do not hear it.

Each layer its own question

Each layer answers its own question. The recurring failure is feeding a surface fact — usually the radical — into a why-it-looks-that-way question.

部首 means section head. Radicals arrange characters into dictionary sections. Paper-dictionary lookup is their one dependable use; a popup dictionary does not make them a second functional analysis. The list still in modern use is the 214 of the 康熙字典, published 1716, rooted in the 字彙 of 1615. Two hundred fourteen is a filing inventory. It is not a study list.

Radical status is per-character and per-standard. Among the six 立-hosts it is radical in only 端 and 章, and in 章 the radical is also non-functional. In 位 the radical is 亻 even though 立 supplies sound and meaning. Radical and functional core are not the same object.

聚 takes sound from 取 and meaning from the three-person cluster. The assigned radical 耳 plays no functional role.

The 錦 inversion makes the diagnostic checkable. A published scholar read 錦 as semantic 金 plus phonetic 帛 and concluded sound components are ineffective. Functionally 金 is sound and 帛 is meaning. Radical-first parsing blinds even capable analysts.

Writing, filing, routing

Taiwan, Hong Kong, and the PRC each maintain a writing standard. Two textbooks can disagree on stroke order with neither being wrong.

戈 is the checkable case. The PRC writes the upper-right dot last. Taiwan writes the bottom 撇 last. Both apply real rules — dots-last against top-before-bottom.

Dot-topped components (言, 宀, 广, 疒) float the top mark in PRC print and touch it in Taiwan print. Native writers and font designers do not always follow official details. Variation in the wild signals nothing alarming.

Writing every stroke correctly under one chosen standard is the prerequisite for correct, good-looking characters. That is a routing rule, not a handwriting course. Stroke-level detail lives in the two books named under Sources; they are optional pointers, not required purchases for this parse.

Filing is standard-relative too. The PRC files simplified 問 under the door 門 (sound). Taiwan files 問 under 口 (semantic). Hong Kong positions on 戈, the dots, 女, and 問 are not used here; they matter only if Hong Kong forms are the target script.

Three questions, three answers.

To understand a character: the functional parts, plus the original sense on Meaning Trees and Original Meanings.

To write it: the governmental standard already adopted.

To look it up on paper: the radical, and no functional conclusions drawn from it.

The method fails where every cut is treated as a finding — the stopping rule has already been left. It does not teach handwriting. Two hundred fourteen radicals will not become a study list by being named. A popup-dictionary reader will rarely need the radical’s one job. The price is one functional parse per character, paid once. After that parse, the functional parts can be named, the radical named separately, and those two answers kept apart.

The same cut is now the thing that stops. Writing and filing go to their own questions. A surface fact is not fed back into why the graph looks that way.

Links into the knowledge base

• How Chinese Characters Work — the hub: this page is the stopping rule the component pages assume.
• Three Attributes of a Character — the triad functional components express.
• Form Components — depicting role.
• Meaning Components — sense-contributing role.
• Sound Components — pronunciation-indicating role; drifted phonetics.
• Empty Components — residue role, a functional verdict, not “below the stopping line.”
• Meaning Trees and Original Meanings — original senses the parse anchors to.

Open Questions

How widespread is standard-relative radical assignment beyond 問?

Hong Kong positions on 戈, dots, 女, and 問 remain unused here.

Sources

Kangxi radicals: 214 section heads popularized by the 康熙字典 (1716), taken from the 字彙 (1615). 部首 is literally “section head.”

Taft & Zhu 1997, and the component-processing neighbourhood around it: functional parts, not arbitrary stroke groups, carry information.

Shu & Anderson 1997: meaning-component awareness in character learning.

Johan Björkstén, Learn to Write Chinese Characters; Harvey Dam, Regular Script Graphemics — optional handwriting pointers, not a curriculum for this page.