Home | Other Papers and Essays

Jim Breen's Japanese Page

Bill Gordon

November 2000


Note:  Jim Breen has made several changes to his web site since this essay was written.  The comments in this essay relate to Jim Breen's Japanese Page and web site as they existed in late October 2000.
Introduction (this page)
Section 1 - Comparison With Other Resources
Section 2 - Site Evaluation Using Standard Criteria
Section 3 - Unique Aspects of Using Web
Conclusion
Bibliography
Jim Breen's Japanese Page (http://www.csse.monash.edu.au/~jwb/japanese.html) offers an abundance of resources related to Japanese language and culture. The web site's main features include an online Japanese-English dictionary server, a frontend interface to the dictionary to translate Japanese text in web pages, an FTP archive of software and files related to primarily the Japanese language, an extensive listing of links to other web sites, and a gallery of ukiyoe prints (Japanese multicolored wood-block prints).

This essay considers the effectiveness of Jim Breen's Japanese Page and his electronic dictionary by evaluating them against standard criteria for web resources and by performing a comparison with other print and electronic resources. The first section compares the features of his electronic dictionary and translation software with print dictionaries, portable electronic dictionaries, and other web dictionaries and translation software. The second part of this essay evaluates his web site by using standard criteria such as authority, scope, organization, content, and reception by target audience. The final section discusses some unique aspects of using the web for the types of resources available on Jim Breen's Japanese Page.

1. Comparison With Other Resources

Japanese-English dictionaries in print have certain useful features that have not yet been reproduced on the web. In addition to print dictionaries, since the early 1990s electronic dictionaries have gained popularity very rapidly because of their speed, portability, and unique features. Besides Breen's dictionary and web page translation aid, other companies and individuals offer online Japanese-English dictionaries and web page translation. The first two parts of this essay section analyze the strengths and weaknesses of print and portable dictionaries in comparison to Breen's web-based dictionary. The third part evaluates Breen's dictionary and translation aid in relation to other web-based dictionaries and translation software.

Print Dictionaries
Publishers have produced many types of Japanese-English print dictionaries, with each one having a specific purpose and its own strengths and weaknesses. However, the major classifications are regular dictionaries, kanji dictionaries, and specialized dictionaries. For each type of dictionary, one or two examples will be compared to the features in Breen's online dictionary.
Regular Dictionaries
Regular dictionaries have two parts, one with Japanese words and English definitions (Japanese-English) and one with English words and Japanese equivalents (English-Japanese). Although some print dictionaries, such as the Random House Japanese-English English-Japanese Dictionary (Nakao 1995), provide only the equivalent word or expression in the other language, the primary strength of a high-quality print dictionary is providing several example sentences and phrases to show the usage of the word. This is especially important for Japanese words, since many words are used with specific particles. For both languages, a one-to-one correspondence of words does not exist, so the lack of usage examples can sometimes lead a person to an incorrect conclusion regarding the use of a word.

Breen's online dictionary provides only word and expression equivalents in the other language, so knowing the word does not mean a person will recognize its appropriate usage. However, many examples of usage in a print dictionary result in much bulk and weight. For example, Kodansha's popular English-Japanese dictionary (1979) and Japanese-English dictionary (1976) take up about 1,550 and 1,250 pages, respectively. Breen's online dictionary has about 60,000 entries, whereas the Kodansha Japanese-English dictionary has a similar number (about 50,000 entries), but Kodansha's dictionary also provides about 60,000 example sentences and phrases. Both of these dictionaries have more entries than the typical commercial dictionary aimed at the beginning or intermediate learner of the Japanese language. For example, the Japanese-English half of the Random House dictionary has about 25,000 entries, but they take up less than 300 pages.

Kanji Dictionaries
As a supplement to regular dictionaries, kanji dictionaries in print have various uses such as providing individual kanji character pronunciations and meanings and giving a list of words (with their meanings) for words beginning with a specific kanji. The print dictionaries provide methods to locate a kanji such as by the total number of strokes, the number of strokes in the radical, or the pronunciation. One of the most authoritative kanji dictionaries for English-speaking users (Nelson 1974) has over 1,000 pages, about 5,000 characters, and almost 70,000 words and compound words. In comparison, Breen's online kanji dictionary has 6,355 characters, and his online dictionary has about 60,000 regular words (not including names and technical terms).

One significant advantage of Breen's web interface to the kanji dictionary is the ability to find words containing a specific kanji either at the beginning of a word or any place within the word, whereas a print dictionary generally only gives words beginning with a particular kanji character. As a second advantage over a print kanji dictionary, a separate interface program allows a user to search the kanji dictionary by clicking on multiple radicals contained within the kanji character. Third, the entries in the online dictionary have about 20 codes that indicate various information about the kanji character (e.g., index number for various classification schemes, frequency-of-use ranking). This advantage however applies to only a few specialists, since most users will never use any of these codes. Breen explains in the dictionary documentation that the file "has information in it which is not much use for people who are not studying and researching Japanese orthography" (Breen, KANJIDIC, 2000).

Specialized Dictionaries
Specialized dictionaries play an important role in communicating between English and Japanese. People who use Japanese often need a dictionary to determine how to read the kanji for a person's family or given name and for names of places such as cities. Breen's online dictionary excels in names by giving users over 160,000 entries, whereas one of the best Japanese-name print dictionaries for English users (O'Neill 1989) has only about 36,000 entries. Breen's dictionary provides great value in determining the pronunciation of names with rare or non-standard readings.

In addition to a dictionary with names, many people require dictionaries for words used in specific fields of work or study since regular dictionaries often do not include many of these technical terms. Breen has tried to add various specialized dictionary files to allow a user to search all of the files at once to try to find technical terms. For example, there are special files for words used in life sciences, aviation, finance, pulp & paper industry, computing and telecommunications, life sciences, business & marketing, law, concrete, and geology (Breen, WWWJDIC, 2000). As a test of the completeness of the business words in Breen's dictionary files, several words were selected from a print dictionary with about 3,000 business terms (Akiyama and Akiyama 1988), but Breen's online dictionary did not contain many of the terms. Based on this limited test, users may want to retain their specialized dictionaries in print, but Breen's effort to gather together specialized terms from various fields should benefit many users.

Electronic Dictionaries
Since the early 1990s, electronic dictionaries have provided almost all of the functionality of regular Japanese-English dictionaries and kanji dictionaries. For example, the popular Canon Wordtank (Model IDX-9500) provides an Japanese-English dictionary, English-Japanese dictionary, and a kanji dictionary all in one machine. The instruction manual (Canon 1993, 64) states the three dictionaries contain a total of about 650,000 entries. For example, the Japanese-English dictionary contains about 90,000 word entries and over 50,000 usage entries. The kanji dictionary contains 6,353 kanji characters, about the same number as Breen's online kanji dictionary. The Canon Wordtank allows the user to jump from one dictionary to another. Other valuable features include a wild card function for searching, the capacity to store up to 500 words for later review, and the capability to insert additional IC cards with specialized dictionaries for business terms and current terms. When traveling away from home or office, most English speakers who speak and read Japanese at an intermediate or advanced level prefer an electronic dictionary to print dictionaries. Many people also prefer an electronic dictionary even when print dictionaries are readily available at home or office.

The speed, portability, and capabilities of electronic dictionaries can lead to the infrequent use of an online dictionary such as Breen's. An electronic dictionary turns on almost instantaneously, whereas it takes time to boot up a PC. Moreover, the limited bandwidth of Internet connections described by Mitchell (1995, 17, 121) five years ago continues to be a problem today for the majority of users, so the slow download times of the Internet can not yet match the almost instantaneous access to information in electronic dictionaries. The lack of portability of most Internet access devices also can discourage users of Breen's online dictionary. However, within the last couple of years, online access through portable devices has grown rapidly. Although few in the US access the Internet through portable devices, more than a third of the Japanese web users access the Internet using wireless technology (primarily with by Internet-enabled cellular phones produced primarily by NTT DoCoMo) (Stevenson 2000). Breen has already developed a cellular phone interface for his dictionary.

Other Web Dictionaries and Translation Software
Several online dictionaries exist on the web, but almost all use Breen's dictionary files (e.g., FreeDict; Friedl 2000; Graf 1999; Heartful Dictionary; Pacific Software Publishing 2000; Palevich 1996; Peterson; Rudick 1997). Chan (2000) states that "all Web-accessible Japanese dictionaries . . . that I've come across seem to use Jim Breen's Edict, Kanjidict, or XJDIC [dictionary files] as the base." Many of these other web sites have developed software to extend the uses and functionality of the dictionary files (e.g., one site uses active HTML to display a translation of Japanese words when the mouse pointer is placed on them) or to provide their own web interface.

Eijiro on the Web (2000) is a dictionary that does not use Breen's dictionary files. However, this dictionary is directed more to native Japanese speakers since furigana (used to determine pronunciation) is not provided for the kanji characters. Based on limited testing, this dictionary appears quite large, with numerous business-related terms not found in Breen's dictionary. A few other web-based Japanese-English dictionaries are directed at native Japanese speakers, but none of those reviewed had the comprehensiveness of Breen's dictionary or Eijiro on the Web.

Besides the ability to look up individual words in the dictionary file, Breen also provides software with a web interface to translate words in Japanese text on web pages or text input by the user. He makes clear in the documentation that the system does not try to translate the entire Japanese text into English. Instead, "it simply attempts to identify the words in the text and to display the translations of those words. The user is expected to know enough Japanese grammar to make sense of the results" (Breen, WWWJDIC, 2000). Breen provides links to two online text translation systems, elingo and Alis Technologies, but he does not "think much of the quality of the results." A search of the web did not identify any other online Japanese-English text translation systems, but several companies and individuals do sell translation software (e.g., Internet Language Company sells bi-directional Japanese translation bundle for $1,395). These commercial translation systems have not been examined in this paper.

In order to test the features of Breen's translation aid and the other two online text translation systems, a typical sentence found in the business section of a large Japanese daily newspaper (Asahi Shimbun October 28, 2000) was selected and then translated into English:

On the US stock market on the 27th, the Dow Jones Industrial Average, which shows the price movements in large company stocks, rose sharply because economic indicators showing a business slowdown were announced, and a sense of relief spread among investors who felt that monetary tightening will end.
Next, the three translation systems were used for this sample Japanese sentence.

Breen's translation aid seemed very easy to use, with the web page shown in the left-hand frame and the Japanese word, pronunciation (hiragana), and definition shown in the right-hand frame. However, the translation aid's weaknesses became readily apparent with this limited test. Breen's system did not translate several of the Japanese compound words (e.g., "stock market," "Dow Jones Industrial Average," "economic indicators") in the sentence, because his dictionary contained only the individual words making up the compound word. However, the dictionary did provide the meanings of some compound words ("monetary tightening," "large company"). The dictionary probably also has difficulty handling one-kanji prefixes and suffixes. For example, the kanji prefix for "US" (bei) is the same as the kanji for "uncooked rice" (kome), but the dictionary only showed the meaning of "uncooked rice." Finally, Breen's translation aid produced a couple of incorrect readings of the text. "Dow" was translated as "down." "Spread" (hirogatta) was translated as the name "Hiroshi," since the software program apparently did not recognize the verb ending (~gatta) of the verb for "spread" even though the Japanese verb hirogaru is in the dictionary.

Both online text translation systems, elingo and Alis Technologies, gave unintelligible results for the sample sentence. The Alis Technologies software provided the following translation:

The dhow industrial stock average by which the fluctuation in prices of the brand of the big enterprise is shown from the extension of a safe feeling that economic indicators which show the business deceleration are announced, and tighter credit ends rises greatly in the US stock market on the 27th.
The elingo system gave the following translation:
As for the American stock market on the 27th, the economic index which shows business deceleration announces When it is done, the tight money ends, the relief impression had to spread Empty and large The Dow Jones industrial stocks average which shows the value movement of brand of enterprise substantially rise.
The results from these two online translation systems suggest that it may be more useful to have only individual words translated, like in Breen's translation aid, rather than getting back a confusing and misleading translation that may be impossible or very difficult to untangle its basic meaning.

Next Section


Introduction | Section 1 | Section 2 | Section 3
Conclusion | Bibliography
"Jim Breen's Japanese Page"
in printer-friendly PDF format:

Home | Top | Other Papers and Essays | About the Author