漢字首尾檢字法概述 bsyx 11/89 10/02

漢字首尾檢字法概述 bsyx 11/89 10/02 9/14

提要

這是「部首檢字法」之一，只在字頭或字尾尋找，部首精減為175個，檢字只依據一個邏輯程序。本文是概述，文末附英文說明，另有詳論兩篇。

漢字是由筆劃組成字元--漢字的單元，再由字元組成漢字。部首可認為是組合頻率較高的字元或字元組，部首大都代表形聲字的意符或會意字的主意符。

漢字迄今尚未有確定的分部法則，也沒有統一的標準部首頒布通行。筆者依據實驗，試在這二篇論文中，提出部首的“定性、定位、定量、定序”的分析報告，以就教於方家。

筆者編部實驗的字集範圍，包括 1986 年在台北出版的<通用漢字標準交換碼>內，常用字 5,401 個，次常用字 7,650 個；及 1988 年 3月在大陸公布的 <現代漢語通用字表> 內 7,000 個；除字形相同不計外，共約 16,000 字。茲摘述論文要旨如次：

1. 定性--漢字是圖形文字，適於見形分部，故提出比較首尾字元結構以定部首的方法；漢字又以“形聲”為其主要特色，傳統的依照意符的選部原則，宜作為除字形結構外另一分部的依据。

2. 定位--漢字部首可定位在字頭或字尾二個位置：先行書寫的字元(部首)，謂之字頭，其位置大都在一字的左方、上方、左上方、右上方及外方。如杜、岩、床、勻、囚，取字頭的字元木、山、广、勹、囗為部首；最後書寫的字元，謂之字尾，其位置大都在一字的右方、下方、左下方、右下方及下外匡如彩、豎、道、斡、凶，取字尾的字元彡、豆、辶、斗、凵為部首，如下表：

字

頭

左方

上方

左上方

右上方

外方

字

尾

右方

下方

左下方

右下方

下外方

杜/木

岩/山

床/广

勻/勹

囚/囗

彩/彡

豎/豆

道/辶

斡/斗

凶/凵

3. 定量--本檢字法只在漢字的字頭或字尾設定部首，部首已加以精減，自首號“一” 部起，至末號“鼻”部止，(就繁體言至“龍”部止），合計部首 175 部。其中 2 劃以上的複筆部首 170 部，單筆部首 5 部。單筆部首都在字頭，即起筆的筆劃：橫H(一)、直I(丨)、撇P(丿)、點T(丶)、彎W(乙) 5 部。

4. 定序--部首照筆劃數及首筆的筆劃性質編定序號，即依筆劃性質：橫(H)、直(I)、撇(P)、點(T)、彎(W)的順序編排。部首有異形者：若筆劃相同，則編為同號，如人/亻編為 2pr，其中2表筆劃數，p表首劃撇，r為部首“人”的代號。若筆劃不同，例如氵/水各編為3tl / 4il。其中 3/4 表筆劃數，t/i 表首劃的性質，即點/直，“l”表部首“水”的代號。同筆劃的部首照“札”字序排列。

此外，筆者擬定分部檢字的一貫邏輯與程序如下，(先選複筆部首，後選單筆部首)：

1. 只有字頭(或字尾)有複筆部首？別無選擇，取之為部首：

例菲：取艸 (“非”不作為部首) 耐：選寸 (“而”不作為部首)

2. 如頭尾都有部首，先就字形結構比較二者在字中所占的權重，取權重大者為部首：

例絮：取糸,不取女擂：取手,不取田 (“手/扌”的權重占1/2，而“田”只占1/4)

3. 如權重不分大小，觀察是否有“主”意符？選主意符為部首，必要時另加次選：

例刖：選刀思：選心寨：正選木，次選宀 (“木”是意符)

4. 如首尾部首權重不分大小，并無主意符，則取字頭部首為正選，字尾為次選：

例古：正選十, 次選口兌：正選八，次選儿

5. 如字頭字尾均無複筆部首，則取該字的首筆為部首：

例久：選丿之：選丶也：選乙亞：選一

依照實際編部的統計，發現依新法分部的結果，與傳統的編排相差不多，新舊法則殊途同歸，表示其可行性大，易為大眾所接受。個人認為就前述 16,000勅漢字言，175勅部首應夠用。漢字在 3 萬字以上的排檢，可能須另增一些部首，但選部的邏輯與規則完全不變。

新編部首簡表 --以首劃橫H、直I、撇P、點T、彎W 為序(合計175個，著色者表繁體)

1H 2H 2I 2P 2T 2W

一丨丿丶乙十厂匚卜冂八人/入勹匕儿几亠冫冖凵卩刀力厶

3H 3I 3P 3T 3W

又廴工士土廾大尢寸弋口囗巾山彳彡夕夂丬/爿广门宀彐/尸嗵

4H 4I 4P

已弓女小子马巛王韦蕒木犬歹车戈瓦菍止攵日/曰贝见气片牛

4T 4W 5H 5I

手毛斤爪父月/肉欠风殳文方火斗户/戶心水爿示石龙业目田

5P 5T 5W 6H 6I 6P

罒皿生矢禾白瓜鸟疒立穴疋皮癶矛耒/耒耳襾页至虍/虎虫 (肉) 缶舌

6T 6W 7H 7I

竹臼血舟色衣羊米聿艮𤠒艸/艹羽糸/纟麦走赤車豆酉豕卤貝見

7P 7T 8H 9H

里足邑/阝身辵/辶釆谷豸角言辛雨齿黾隹阜/阝金鱼門革頁骨香

10 11 12 13+

鬼食風音首韋鬲髟馬鬥黃麥鹵鳥魚麻鹿鼎黑鼓黽鼠鼻齒龍

漢字首尾分部新法圖解

字頭或字尾 N 取首劃為部首

有複筆部首？

例- 氏：取丿巴：取乙

Y 更：取一之：取丶

字頭或字尾 Y 只有一種選擇

只一種部首？

例- 菲：選艸令：選人忽：取心

N 嗣：選口耐：選寸麼：取麻

頭尾部首

比較,其所占權重 N 不相當,取權重較大者為部首*

是否相當？例- 絮：取糸,不取女擂：取手,不取田

Y 二者權重相當

是否有 Y 有主意符,取之為正選,其他為次選(必要時)

主意符?

例- 刖：選刀 (不必次選) 碩：正選頁, 次選石

| 思：選心 (不必次選) 擲：選手/扌

N 無(主)意符,取字頭為正選,字尾為次選(必要時)

例- 員：正選口, 次選貝古：正選十, 次選口

* 由漢字的結構及部首在字中的布局，可由直觀判別部首在該字中所占的權重。部首所占位勢較強者，則其權重較大。若為形聲字，可假定意符的權重與聲符的權重相當：

例得：彳之權重較寸為大，彳占 1/2 而寸只占 1/4

庶：广之權重較灬為大，因广据有二邊，且含灬

若能了解漢字結構，則權重的判斷更加容易：

例捌：意符手與聲符別相當，故手重于刂，選手

恚：心與圭相當，心的權重大于土，選心

如為會意字，則取 "主意符" 為部首：

例-- 尘：小土為尘，土為名詞，乃主意符，取之為正選部首

ＮＥＷＲＡＤＩＣＡＬＳＹＳＴＥＭＯＦＣＨＩＮＥＳＥＣＨＡＲＡＴＥＲＳ 3/97 10/02

Chinese characters in a dictionary are usually indexed by radicals (部首). The number of radicals varies from 200 to 214 according to the existing radical system. However, as of today, there are no standard rules provided for deciding each word's radical. Most of the Chinese dictionary users have to memorize the words' radicals or sometimes can only try several times before they locate the right words.

The new radical system has only 175 roots in total which can be applied both complex- and simplified-form characters. The radical is either the beginning or ending part of the word. For example, 艹 , the beginning part of word 菲, is the radical and 殳, the ending part of 穀 , is the radical (instead of 禾). When the beginning or ending part of a word does not belong to any of roots, the first stroke of this word is then its radical. For example," 一" is the radical of word 互 , and "丿" is the radical of word 乏.

When the beginning and ending parts of a word are both of the 175 roots, the part placing major position of the character is selected as the word's radical. For instance, 月 instead of 馬 is selected as the radical for word 騰；門 instead of 耳 is selected as the radical of word 聞. In case such two parts stand equal space of a word, either can be the radical of the word. For example, 克 has two radicals (十 and 儿) and select 儿 as the principle radical and 十 for sorting purpose only.

The 175 roots of the New Radical System are arranged in the sequence of stroke numbers. The radicals with equal strokes are listed according to their first stroke categories with English alphabet sequence. i.e. H (heng 一), I (zhi 丨), P (Pie 丿), T (dian) and W (wan 乙).

The New Radical System assigns codes to each of the 175 roots indicating its stroke number, category and a specific English alphabet relating to its pronunciation. Every Chinese word can be coded by combining its radical's codes with the number of its strokes, thus facilitating sorting or indexing of Chinese words by the computer. For example, 人/亻is coded as 2pr

where‘2’is the stroke number of 人,‘p’is the category of first stroke‘pie’of 人 and‘r’ relating to its pronunciation.

A total of 16,000 words including 13,051 standard complexes-form issued in 1986 in Taiwan, and 7,000 standard simplified-form words (of which about 4500 are the same as complexes-form) issued in 1988 in Peking have been indexed by the author according to the new radical system. The new system is deemed applicable to all Chinese words perhaps with a few more radicals added for more words than 30,000. Anyhow, the basic principle for selecting radicals will be no change, since most of Chinese characters are of similar structures.

In short, the new radical system has fewer roots (175 instead of 214) applicable to both complex-and simplified-form Chinese characters and the radical in each word can be clearly determined (either beginning or ending part of the word). It is believed the new Radical System will help to substantially reduce the time for Chinese dictionary looking up.

The following Table indicates the new 175 Radicals arranged by their stroke numbers, positions and first stroke categories (H I P T W).

175 ＮEW ＲADICAL ＴABLE (新部首表)

B：Beginning(字头) E：Ending(字尾) =：ditto to left(同左) [ ]= complex-form in red (繁体用部首)

H= horizontal(横) I= vertical(直) P= diagonal(撇) T= dot(点) W= bending (弯)

STROKE	B E	B E	B E	B E	B E	B E	B E	B E	B E	B E
１劃	一	︱	丿	丶	乙
２劃 H/I	十 =	厂	匚	卜 =	冂
P	八 =	人亻	勹	匕 =	儿	几 =
T	亠	冫	冖
W	凵	卩	刀刂	力 =	厶 =	又 =	廴
３劃 H	工 =	士	土 =	廾	大 =	尢	寸	弋
I	口 =	囗	巾 =	山 =
P	彳	彡	夕 =	夂 =
T	丬	广	门	宀
W	彐 =	尸	己 =	弓 =	女 =	小 =	子 =	马 =	巛
４劃 H	王玉	韦 =	蕒	木 =	犭犬	歹	车 =	戈	瓦	菍
I	止 =	攴攵	日 =	贝 =	见
P	牛 =风 =	扌手殳	毛 =	气	片	斤	爪	父	月肉	欠
T	文 =	方	火灬	斗	户戶	忄心
W	氵水	[爿]
５劃 H	礻示	石 =	龙 =
I	业	目 =	田 =	网/罒	皿
P	生 =	矢	禾 =	白 =	瓜 =	鸟 =
T/W	疒	立 =	穴	疋 =	皮 =	癶	矛
６劃 H	耒	耳 =	襾	页	至
I	虍虎	虫 =	(肉)
P	缶 =	舌	竹	臼 =	血	舟	色
T	衤衣	羊 =	米 =
W	聿 =	艮 =	艸/艹	羽 =	糸 =
７劃 H	麦	走	赤	[車] =	豆 =	酉 =	豕 =
I	卤	[貝] =	[見]	里 =	足 =	邑
P	身	辵	釆	谷 =	豸	角 =
T	言讠 =	辛 =
８劃	雨	齿 =	黾	隹 =	阜	金钅 =	鱼 =	[門]
９劃	革 =	[頁]	骨	香 =	鬼 =	食饣 =	[風] =	音 =	首 =	[韋] =
１０	鬲 =	髟	[馬] =	鬥
１１	黄 =	[麥]	[鹵]	[鳥] =	[魚] =	麻	鹿 =
１２	鼎	黑 =
１３+	鼓 =	[黽]	鼠	鼻	[齒] =	[龍] =

Logical Steps for Picking-up Radical of a Chinese Character

(By Word-Head or-Tail Method)

Keywords-

Word-Head(字頭):the first part (component) of a word to be written, e.g. 日 in word 昭

Word-Tail(字尾):the last part (component) of a word to be written, e.g. 糸 in word 絮

Radical (部首): A radical can be selected either in the word-head or word-tail as far as the word structure concerned. The radical usually

represents the meaning part of a phonetic- complex (形聲字) or major meaning part of a logical- aggregates (會意字).

Step 1-- Is there a Multi-stroke-Radical in the Head or Tail of the word?

No, Pick-up the first stroke of the word as Radical. e.g. 久：select[丿], 之：[丶],也：[乙], 亞：[一]

Yes, Follow step 2

Step 2 --Is there only one Multi-stroke-Radical in the head or the tail?

Yes, Pick it up as Radical of the word. e.g. 菲：select[艸], 耙：[耒], 嗣：[口], 耐：[寸]

No, Both head & tail having radicals, follow step 3

Step 3 --Compare the Weight/significance of the Head’s and Tail’s Radical, Pick-up a significant ones as a Radical. e.g. 絮：select[糸] instead of [女], 擂：[扌] instead of [田]

Step4 --If both Head & Tail Radicals’ weight is same & one Radical’s meaning related to the word, Pick-up the meaning-related one as radical. e.g. 刖：select[刀], 急：[心], 碩：[頁], 擲：[扌],彬：[彡]

Step5 --If both Radicals’ weight is same & neither one related to the word meaning, Pick-up the Head’s one as the Radical of the word. e.g. 沓：select[水]; 古：select [十] & [口] as reference.