JAPANESE 8500 - Japanese Bibliography and Research Methods

Problems and procedures in the use of Japanese bibliographies and other reference materials, including electronic sources.


Japanese lexicons (語彙集) and Japanese dictionaries (国語辞典) have their entries organized according to a few well-understood principles. In modern dictionaries, the operating principle is pronunciation, and the ordering is the 五十音順 (gozyuu'onzyun: See below). This makes the dictionary itself a collection of definitions, indexed according to the lexical entries to which they correspond.

In contrast, Japanese thesauri (類似語字典) will typically have their entries organized by semantic categories, and will include a separate index of lexical items based on pronunciation (to help you find a category by looking for one of its more familiar members).

Chinese-Japanese character dictionaries (漢和辞典) typically have separate indexes of several types to allow you to find a given 漢字 in several different ways. Among these is an index based on pronunciation, but this is usually not the principle by which the entries themselves are ordered. Most 漢和辞典 organize their entries based on 部首.

One rare exception which lists its main entries by 音・訓読み and includes a 画数索引 is this:

  • Hikkei Kanji jiten. Sanseidō henshūjo (hen). Tokyo: Sanseidō shoten, 1992
必携漢字辞典. 三省堂編修所(編). 東京: 三省堂, 1992.
OCLC: 30830351
LC: PL677.5

In the following discussion we concentrate on the skills necessary for using a more conventional 漢和字典.


The formation of 漢字 as the composition of radicals (or 部首), is known as 偏旁 (へんぼう). (For details, see "More on 部首" to the right)

The class of 漢字 is defined in such a way that all 漢字 incorporate at least one 部首. Among the 部首 in a given 漢字, there must be exactly one principle 部首 (according to which that 漢字 is classified). In many cases, a 部首 in isolation constitutes a 漢字 as well.

The Chinese dictionary 説文解字 was the first to order entries in a dictionary according to the principle 部首 (540 of them in all, themselves partially ordered by stroke count). Several subsequent dictionaries such as the 字彙 and the 正字通 based their ordering on a smaller set of 部首 numbering 214. The system most commonly used today is sometimes referred to as 康煕字典順 (こうきじてんじゅん), after the 康煕字典, which follows this principle.

While you can expect to find at least three indexes in a comprehensive 漢和辞典, for almost all of them the organizing principle for the main entries themselves is based on 部首, and for any given 漢字 a unique main entry will be classified under its principle 部首.

For 漢字 incorporating more than one 部首, some 漢和辞典 will give a brief entry under the secondary 部首 redirecting the user to the appropriate section.

In 部首 indexes, radicals are partialy ordered according to stroke count (画数). 

In a typical 漢和辞典, you will probably find an index of 部首 either printed on the front inside cover and (front paste-down + front flyleaf = front endsheet), or on the back inside cover (back paste-down + back flyleaf = back endsheet), or as an index section.

Finding a 漢字 by referring to the 部首 is usually a matter of identifying the 部首 by which the 漢字 is classified, finding that 部首 in the index, and referring to the part of the dictionary indicated in the index. The entries for the various 漢字 classified under that 部首 are typically ordered by stroke-count. The compound lexical items listed under each main 漢字 entry are listed sometimes by pronunciation, sometimes by the stroke count of the second character.

This method of indexing comes with some complications. As noted above, although a 漢和辞典 may only have one full entry per character, many 漢字 are composed of more than one 部首. Deciding which 部首 is used for classification is not alway straightforward.

Furthermore, some 部首 have the appearance of being analyzable into combinations of other 部首. 

To make matters worse, some 部首 are based on simplifications of more complex characters, so the form of the 部首 does not always give a direct indication of the stroke count.

Thus you need to know about the system of 部首 as a whole before you can use the index effectively.

For example, you might want to look for a 漢字 that incorporates the common radical 艹 , which is written with three strokes. You might be surprised that in many 漢和辞典 you will not find 艹 among the radicals listed in the 三画 section. This is because 艹 is a simplification of the older form 艸, which is written with six strokes (and incidentally has another variant ++ of four strokes, as well). Look in the 六画 section and you will find the page number for the appropriate section of the main entries where you can start looking for 漢字 that incorporate the radical 艹.

Here are a few 部首 where this problem shows up:

Radicals with variants of different stroke numbers
radical stroke count variant 1 stroke count variant 2 stroke count note
阝(右) 2 7      
阝(左) 2 8      
2/3 Shinnnyou.gif 3/4 7 (see しんにょうの点)
3 4 Sitagokoro.gif (下) 4  
3 4      
3 Kusa-4.gif 4 6  
3 4 氺 (下) 5 Sitamizu.gif (下)
3 4      
3 4      
4 6      
4 5      
罒 (上) 5 6      
4 5      
6 5      
11 7      
12 5      

In the following cases, as above, the forms differ while the classification is the same. But there is no difference in stroke number in the examples below.

Radicals with variants having identical stroke numbers
radical stroke count variant 1 variant 2
2 Hitogasira.gif (上)  
3 Naogasira.gif (上)  
灬 (下) 4  
4 爫(上)  

There are many more variations.

One of the reasons why it is crucial to recognize variations in the forms of radicals is that a change in stroke count may not only affect the place in the order of radicals, but may also affect the place in the order of characters in the 画数 index. Which brings us to the next section.


As a general principle, 画数 is used to 1) partially order 部首 in the 部首索引; 2) partially order the 漢字 classified under a given 部首; and 3) partially order the 漢字 arranged in the 画数索引.


Some things to pay attention to:

  1. What is the order of voiced stops?
  2. What is the order of long vowels in loan words?
  3. What is the order of palatalized morae?


Here is the order:










Warning: 古語辞典 (and some older 国語辞典) order their entries according to the 五十音順 of 歴史的仮名遣い. Not only does this include various symbols and combinations of symbols not commonly used in Modern Japanese, but many cognates to familiar words may also have different "spellings." A well-known example of differences following the language reforms after WWII is the "spelling" of the word 蝶々 as ちょうちょう (originally てふてふ). Naturally, the 五十音順 indexes of some older 漢和辞典 also exhibit these characteristics.

Modern Japanese 五十音順 is organized phonologically to a certain extent: The combination of five vocalic values


with 10 consonantal values


gives you (in theory) 50 syllables.

Only 45 symbols are actually used in the indexing system of a modern 国語辞典. Here is the order:


The sounds [yi], [ye], and [wu] aren't distinguished in Japanese, and so don't have corresponding symbols.

The symbols corresponding to sounds [wi] and [we] are, respectively, ゐ and ゑ. But these are no longer used in the modern writing system.

The symbol corresponding to syllabic [N] is used in the modern writing system: ん. But as it never occurs word-initially, it has no place in the indexing system, except in a 逆引き辞典.

Alphabetical order

Here is the order:


Where to find the word in question depends on how it is romanized.