Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

Available

Available online
online at
at www.sciencedirect.com
www.sciencedirect.com

ScienceDirect
Available online at www.sciencedirect.com
Procedia
Procedia Computer
Computer Science
Science 00
00 (2017)
(2017) 000–000
000–000
www.elsevier.com/locate/procedia
ScienceDirect www.elsevier.com/locate/procedia

Procedia Computer Science 112 (2017) 1345–1352

International Conference on Knowledge Based and Intelligent Information and Engineering


Systems, KES2017, 6-8 September 2017, Marseille, France

Analysis of Dialogues Difficulty in Anime


Comparing with JLPT Listening Tests
Shan Junjieaa, Yoko Nishihara, Ryosuke Yamanishi, Junichi Fukumotobb*
a
aGraduate
Graduate School
School of
of Information
Information Science
Science and
and Engineering,
Engineering, Ritsumeikan
Ritsumeikan University
University
b
bCollege of Information Science and Engineering, Ritsumeikan University
College of Information Science and Engineering, Ritsumeikan University

Abstract
Abstract

More
More andand more
more people
people choose
choose toto learn
learn Japanese
Japanese asas aa foreign
foreign language
language (JFL)
(JFL) in
in the
the world.
world. TheThe traditional
traditional teaching
teaching materials
materials have
have
focused
focused onon training
training for
for the
the skills
skills of
of writing
writing and
and reading
reading inin Japanese.
Japanese. However,
However, it it seems
seems that
that the
the JFL
JFL learners
learners would
would like
like to
to train
train
their
their speaking
speaking and
and listening
listening rather
rather than
than writing
writing and
and reading
reading because
because the
the speaking
speaking and and listening
listening are
are necessary
necessary to to make
make
conversations.
conversations. We We use
use Japanese
Japanese animation
animation which
which isis called
called “Anime”
“Anime” in in the
the world
world as as aa new
new teaching
teaching material
material ofof Japanese.
Japanese. Anime
Anime
has
has aa huge
huge amount
amount ofof scenes
scenes in
in which
which characters
characters speak
speak with
with their
their voice
voice in
in the
the standard
standard andand clear
clear pronunciations.
pronunciations. It It means
means that
that
Anime
Anime hashas many
many good
good examples
examples of of Japanese
Japanese dialogues
dialogues for
for JFL
JFL learners
learners to
to improve
improve their
their speaking
speaking andand listening.
listening. The
The goal
goal of
of our
our
research
research isis to
to propose
propose aa newnew method
method to to classify
classify dialogues
dialogues in in Anime
Anime depending
depending on on the
the dialogues’
dialogues’ degrees
degrees ofof difficulty.
difficulty. In
In this
this
paper,
paper, wewe analyzed
analyzed the
the words
words and
and expressions
expressions used
used inin the
the Anime
Anime dialogues
dialogues and
and the
the scripts
scripts ofof listening
listening tests
tests of
of the
the previous
previous
Japanese
Japanese Language
Language Proficiency
Proficiency Tests
Tests (JLPT).
(JLPT). Each
Each word
word and
and expression
expression were
were classified
classified into
into different
different Japanese
Japanese language
language levels.
levels.
We
We used
used the
the analysis
analysis results
results of
of JLPT
JLPT as as aa standard
standard reference
reference to to be
be compared
compared with
with Anime
Anime dialogues.
dialogues. The
The analysis
analysis results
results showed
showed
that
that the
the genres
genres ofof Anime
Anime had had effects
effects onon using
using words
words and
and expressions.
expressions. We We also
also found
found that
that the
the scripts
scripts of
of listening
listening tests
tests of
of high-
high-
level
level included
included many
many high-level
high-level words
words and
and expressions.
expressions.
©
© 2017
2017 The
The Authors.
Authors. Published
Published by by Elsevier
Elsevier B.V.
B.V.
© 2017 The under
Peer-review Authors. Published by
responsibility of Elsevier
KES B.V.
International.
Peer-review under responsibility of KES International.
Peer-review under responsibility of KES International
Keywords:
Keywords: Anime,
Anime, Japanese
Japanese language
language learning
learning support,
support, Words
Words and
and expressions
expressions of
of Japanese,
Japanese, Speaking
Speaking and
and listening
listening abilities
abilities

*
* Corresponding
Corresponding author.
author.
E-mail
E-mail address:
address: gr0260hr@ed.ritsumei.ac.jp
gr0260hr@ed.ritsumei.ac.jp

1877-0509
1877-0509 ©© 2017
2017 The
The Authors.
Authors. Published
Published by
by Elsevier
Elsevier B.V.
B.V.
Peer-review under
Peer-review under responsibility
responsibility of
of KES
KES International.
International.

1877-0509 © 2017 The Authors. Published by Elsevier B.V.


Peer-review under responsibility of KES International
10.1016/j.procs.2017.08.022
1346 Shan Junjie et al. / Procedia Computer Science 112 (2017) 1345–1352
Jungie Shan et al / Procedia Computer Science 00 (2017) 000–000

1. Introduction

Japanese animation called Anime in the world is the most famous culture of Japan. There are many lovers who
enjoy watching Anime. Anime festivals have been held many times not only in Japan but also in western countries,
e.g., France and Canada. Foreign people respects Anime and its creators just like Hollywood movies and their
creators. Foreign people may not know Japanese enough so they might not directly understand the speaking Japanese
in Anime. They read subtitles on Anime which are translated into their own native language. Now they are enjoying
Anime by reading the subtitles while listening to Japanese as just sounds.
Anime draws various scenes. Some of Anime draw battle scenes, the others draw daily lives. In the scenes,
characters speak with their voices. The voices are given by voice actors who have be trained as the specialists of
Japanese speaking for many years even though almost all voice actors are native Japanese. The characters make
dialogues in the Anime scenes. That is, Anime has a huge amount of Japanese dialogues with the standard and clear
pronunciations.
Anime could be a good material for studying Japanese. In particular, speaking and listening in conversations may
be trained well by using Anime. Through watching Anime, some of the foreign people have acquired listening and
speaking ability for Japanese [7,8]. More and more people learn Japanese as a foreign language (JFL). According to
the Japanese Foundation, there were about 3.6 millions of people who studied Japanese abroad at Japanese
educational organizations in 2015, distributed in 137 countries [1]. The JFL learners are requiring good materials for
learning Japanese, especially for their speaking and listening in Japanese.
There are two problems usually be mentioned in educational practices' feedback. The first problem is inadequate
teaching materials have been used. Teachers who teach Japanese to foreign people have not enough time to update
their teaching materials. Therefore, the teaching materials are too old to teach Japanese used nowadays. Traditional
teaching materials are becoming out-of-date and more and more difficult to satisfy with the JFL learners’ growing
demands.
The second problem is that speaking and listening abilities have been ignored. Usually, Japanese teaching has
focused on almost only reading and writing abilities. As the practical application skills, the speaking and listening
abilities might not be treated seriously in the teaching practices. The speaking and listening are more important than
others in studying foreign languages. A research conducted through JFL learners in the U.S. showed that the
students' confidence would be much enhanced when they noticed their own improvement of listening and speaking,
i.e. conversation abilities [2].
To solve the above two problems, we try to use Anime as a teaching material for Japanese. Anime contains
various dialogues. The dialogues involve nearly every aspect of conversations. The dialogues of Anime are usually
spoken with standard and clear pronunciations. And, scenes of Anime have so attractive and interesting so JFL
learners can continue to practice their conversation abilities using Anime as the learning material.
In order to utilize the large quantities of Anime, we propose a method to classify dialogues in Anime and estimate
the dialogues’ difficulty. In this paper, we analyze words and expressions used in the Anime dialogues and the
scripts of listening tests in the previous Japanese Language Proficiency Tests (JLPT). Each word and expression are
classified into different Japanese language levels. Then, the analysis results of JLPT are used as a standard reference
for comparing with Anime dialogues.

2. Related Works

As one of the most famous Japanese cultures, there are many types of research surround Anime. Most of those
are about using Anime to help Japanese language study especially in the culture dimension [3, 4].
Since 2009, Yazaki’s series papers titled “Anime-de Nihongo: Japanese using Anime” have introduced his
research about using Anime in teaching Japanese [5]. In these teaching practices, the author investigated the
applications of Anime in developing new teaching methods and explored the effects of different types of student
groups respectively. Through all the practices had got positive results, the focus of his practices was all about
developing new “Teaching Strategies”. Development of new teaching materials is also required for JFL learners
abroad. We try to develop new learning materials in our study.
Shan Junjie et al. / Procedia Computer Science 112 (2017) 1345–1352 1347
Jungie Shan et al / Procedia Computer Science 00 (2017) 000–000

Table 1. Numbers of Words in Each Level of JLPT. There are four levels. L4 is the primary. L1 is the most difficult
Level L4 L3 L2 L1
Number of Words 738 685 3636 2962

Table 2. Examples of word in each level of JLPT


Level Example of word
青い(blue), 行く(go), 後ろ(back), 映画館(movie theater), 同じ
L4
(same), 学校(school), 可愛い(cute), 果物(fruit)
挨拶(greeting), 意見(opinion), 美しい(beautiful), 選ぶ(select), 贈
L3
り物(gift), 会場(meeting place), 季節(season), 比べる(compare)
愛情(love), 医師(doctor), 失う(lose), 液体(liquid), 応用
L2
(application), 片道(one way), 記憶(memory), 苦情(complaint)
愛想(friendly), 遺跡(remains), 団扇(Japanese fan), 英雄(hero),
L1
疎か(negligent), 介護(nursing), 禁じる(forbid), 空腹(hungry)
The previous researches in the field of assisting Japanese E-Learning systems were usually took sight in written-
form, say, vocabulary retention or reading comprehension. Such as extract the words or phrases from the sentences,
and help JFL learners with remember the vocabularies.
Hazelbeck et al. developed a system for extracting Japanese words from sentences more accurately and separate
them into different JLPT levels[9]. They have tested two kinds of corpus in comparing the word coverage rates. The
system even could measure the JFL learners' degrees through their learning processes and display the effective
explanation ranked by the degree of difficulty.
A similar attempt using mobile-based Email for helping JFL learners with learning Japanese Kanji also adopted
the same strategy[10]. LI et al. divided difficulty levels based on the JLPT vocabulary lists to suit each leaner's'
cognitive level. Then they support Japanese Kanji learning through mobile-based Email and SMS individually.
A latest study also investigated the Japanese-language grammar in learning assistant[11]. Zhe et al. extracted the
sentence patterns and vocabularies from Japanese sentences in each level, and utilized the construction relation
diagram in Japanese grammar teaching. They use their system as an assistance with the traditional textbooks.
We have also already made statistics of Anime dialogues' level using the Japanese expressions with different
genres of Anime [6]. Through the analysis, we found that in different genres of Anime the level of Japanese words
and expressions used in Anime were different respectively. Moreover, the level of dialogue might be different
according to the genres of Anime. All JFL learners’ levels of Japanese are not equal, but different. To use Anime as
a learning material, the dialogues of Anime should be classified as their levels of Japanese.
How should we subdivide the dialogues? To solve the problem, in this paper, we analyze the level of dialogue of
Anime by comparing sentences of Japanese listening test of Japanese Language Proficiency.

3. Analysis Method

In order to estimate dialogues’ difficulty in Anime, first of all, we make subtitle texts of Anime dialogues. Then
we analyze the subtitles based on two aspects: words and expressions.
According to JLPT, there are four levels of word and three levels of expression determined by JLPT in order to
check JFL learners’ learning degrees, Level 4 (L4) to Level 1 (L1). L4 is the primary though L1 is the most difficult.
Each level has its own words and expressions that the learners need to study.

3.1. Japanese Words of JLPT

JLPT has formulated different word lists for each level. JFL learners need to study those in the corresponding
levels. The numbers of words in each level are shown in Table 1. Examples of the word in each level are shown in
Table 2. The number of words in L2 is the largest though the number in L3 is the smallest. There is a gap between
L2 and L3 in the numbers of words. It may be caused because the words of L2 are practical used in daily lives
though the words of L3 are basic.
Jungie Shan et al / Procedia Computer Science 00 (2017) 000–000
1348 Shan Junjie et al. / Procedia Computer Science 112 (2017) 1345–1352

Table 3. Number of expressions in each level of JLPT. There are three levels. L3 is the primary, L1 is the most difficult.
Level L3 L2 L1
Number of expressions 108 177 105

Table 4. Examples of expression in each level of JLPT


Level Example of word
L3 ~く\になる(become)、~だろう(may)、~ている(be …ing)
L2 ~かのようだ(seem to)、~恐れがある(It has an afraid )、~にしても(even if)
L1 ~にあって(at)、~ながらも(while)、~かたわら(besides)

Table 5. Information of Anime subtitle texts


Total number of rows in
Anime Name Number of episode
subtitle texts
Hyoka 23 9221
Fate Zero 24 7394
K-ON 2 24 10405

The procedures of our analyzing words in Anime dialogues are:


1) We use a morpheme analysis soft (Igo) to separate the sentence in Anime dialogues into words. Japanese
does not have any space between words, thus the morpheme analysis is necessary to understand which
word is used in the dialogue.
2) We use the separated words to match with the identical one from the JLPT word lists and count up the
numbers of words every level.
3) We make the count up result with each Anime dialogue.

3.2. Japanese Expressions of JLPT

JLPT also has formulated different expression lists for each level. Unlike with those for the words, the study of
Japanese expressions is mainly started from L3, not from L4 JLPT has expression lists of L1, L2, and L3. The
numbers of expressions in each level are shown in Table 3. Examples of the expression in each level are shown in
Table 4.
Japanese expressions are also important to get Japanese speaking and listening. In a realistic Japanese dialogue,
words and expressions that native people use depend on the dialogue situation. For example, when the dialogue
happens between friends such as classmates, the words and expressions people use would be more easy than the one
in the service industry. Thus, both words and expressions are studied to cover the possible distinctions of each
Anime. The procedures we analyze the expressions are:
1) We make the lists of Japanese expressions in each level with regular expression's format.
2) We use the regular expressions list to match with Anime dialogues and make count up results.

3.3. Analyzed Subtitle Texts of Anime Dialogues

The information of subtitle texts in this study is shown in Table 5. We use three stories of Anime because the
episode numbers of them are similar, but the genres of the story are different. The “Hyoka” is mainly about a
mystery story, “Fate Zero” includes more battle scenes, and “K-ON” is completely a story in a school life. We used
three stories of Anime with different genres because we would like to find effects of Anime genre for words and
expressions used in the dialogues.
Shan Junjie et al. / Procedia Computer Science 112 (2017) 1345–1352 1349
Jungie Shan et al / Procedia Computer Science 00 (2017) 000–000

Table 6. Statistics of words and expressions in Hyoka

Episode L4 L3 L2 L1 Non
Words 888 274 461 238 378
1
Expressions none 276 135 23 -
Words 843 203 415 157 283
2
Expressions none 267 109 15 -
Words 888 229 379 160 269
3
Expressions none 280 108 13 -
Words 839 241 452 240 292
4
Expressions none 250 121 13 -
Words 912 260 410 230 369
5
Expressions none 294 116 10 -
Rate of Words 42.6% 12.2% 20.7% 9.15% 15.1%
Rate of Expressions none 68% 28% 4% -

3.4. Previous JLPT Listening Tests

As the standard comparison reference, we use the real JLPT tests in the past. The JLPT tests consist of three parts:
Listening, Grammar and Reading. The latter two mainly examine writing abilities such as vocabulary accumulation
and grammar usage. Because of the goal of our research, that is to support JFL learners to improve their speaking
and listening abilities, we mostly pay attentions to the Listening part which has a lot of different dialogue scenes and
mainly examine the learners' listening capabilities.
First, we collect the original texts of every levels' listening tests for five years. Second, we make statistics of the
words and expressions appeared in these texts with levels. Then, we use the statistics from the same aspect of Anime
dialogues to compare with the realistic listening tests. Through these two aspects of analysis, we could evaluate the
degrees of difficulty in different Anime dialogues.

4. Discussion

First, we have counted up all the words and expressions appeared in Anime dialogues with different levels,
episode by episode. Table 6 is a part of results of Anime “Hyoka” about front 5 episodes in its dialogues. The Non
of words means the number of nouns that could not match any word in all levels’ list in the dialogues. They maybe
the names of characters, proper nouns or some words beyond the range of JLPT test. The averaged rate of Non
words was 15.1%. We try to estimate the words levels by other methods.
Through the results in Table 6, we could found both words and expressions appeared in Anime dialogues were
concentrated in low degrees, especially in L3 and L4 levels. The total of the average rate of words in L4 and in L3
was 54.8% (=42.6%+12.2%). We suspect that the results were obtained because of three reasons below:
(1) The words and expressions in high levels are more used with written abilities than conversation abilities. As
conversation-based Anime, dialogues in Anime would often include the oral usage of words or expressions in it. For
example, the content in news report in the following sentence:
「もっとも、国が高等教育にかかる費用すべてを負担した場合、年間5兆円かかるとの試算がある。
(Although, they estimate that five trillion yen per year are necessary for the high education if the government
paid all of those fees.) 」
The same meaning of the sentence when people speak to each other may become:
「でも、もし国が高等教育にかかる費用を全部負担したら、一年で5兆円くらいかかると思う。
(But, if the governments pay the all of the fees for the high education, I think it takes five trillion for each year)」
The「場合 (although)」and 「試算(estimate)」are high-level words that would be usually used in written
sentences. Those words may be not often used in conversations. When in conversation the words would change to
easier ones: 「したら(if)」and「思う(think)」.
1350 Shan Junjie et al. / Procedia Computer Science 112 (2017) 1345–1352
Jungie Shan et al / Procedia Computer Science 00 (2017) 000–000

Table 7. Statistics of words and expressions in Fate Zero

Episode L4 L3 L2 L1 Non
1 Words 1306 476 814 387 775
Expressions none 411 268 38 -
2 Words 513 145 300 119 223
Expressions none 165 85 12 -
3 Words 600 184 367 167 270
Expressions none 208 131 32 -
4 Words 383 115 271 140 265
Expressions none 136 90 22 -
5 Words 381 110 259 126 289
Expressions none 117 94 32 -
Rate of Words 35.2% 11.2% 22.0% 11.2% 20.1%
Rate of Expressions none 56% 37% 7% -

Table 8. Statistics of words and expressions in K-ON 2

Episode L4 L3 L2 L1 Non
1 Words 601 219 248 86 216
Expressions none 187 33 14 -
2 Words 796 229 318 106 255
Expressions none 241 85 15 -
3 Words 637 219 313 97 231
Expressions none 224 59 8 -
4 Words 690 159 295 90 206
Expressions none 170 56 11 -
5 Words 648 193 216 49 148
Expressions none 197 38 9 -
Rate of Words 46.9% 14.2% 18.7% 5.2% 14.8%
Rate of Expressions none 77% 20% 3% -

(2) The dialogues of Anime used in this analysis were normally about the daily life conversations, such as
between friends and classmates. The usage of words and expressions in this type of conversations also would not be
very complicated. Here is an example of dialogue between a tourist guide and customers:
「お客様、右手をご覧ください。マルコポーロの生家を復元した建物でございます。(On your right
side, that is the replica of the building where Marco Paulo lived.)」
But the dialogue between friends in “K-ON 2” usually likes:
「見てみて!すごいネジ!(Hey, look! What a great screw.)」
The first sentence is more complicated than the second sentence because the first sentence is composed for
people who may not share all contexts. The second sentence is less complicated because people in the conversation
would share all contexts so they can understand the intention and the meaning even if information is missed in the
second sentence.
(3) The genre of Anime also seemed to cause the different emphases in dialogues’ choosing. An Anime which
tells a general daily life story would often have a relaxed conversation environment, so that the dialogues in it also
will be easier. Besides, an Anime which includes a complex background may need more contents in helping with
explaining the story. For example, when the character explained what the exact background story is in “Fate Zero”,
it says:
「にわかには信じがたい話だとは思うが、あらゆる時代、あらゆる国の英雄が現代に蘇り。
(You do not believe this story though, from every era, every contour the all of the heroes are revived in the
present day.)」
Shan Junjie et al. / Procedia Computer Science 112 (2017) 1345–1352 1351
Jungie Shan et al / Procedia Computer Science 00 (2017) 000–000

Table 9. Line numbers of previous JLPT listening tests’ script


Total line number of listening tests’ script each year
Level
2003 2004 2005 2006 2007
L4 244 186 181 180 185
L3 312 288 292 275 286
L2 388 370 380 364 376
L1 404 405 374 424 383

Table 10. Averages of words in listening tests from L1 to L4.


Level L4 L3 L2 L1 Non
L4 94.17% 1.41% 0.65% 0% 3.75%
L3 79.55% 12.6% 1.79% 0.15% 5.89%
L2 66.74% 12.77% 11.56% 0.53% 8.38%
L1 59.27% 13.64% 15.27% 3.78% 8.01%

The sentences like this are rarely used in daily conversations.


As the comparison, Table 7 and Table 8 show a part of results of Anime “Fate Zero” and “K-ON 2” with front
five episodes.
The reason why numbers of words and expressions are much larger in the first episode of Fate Zero is mainly
because the duration of the first episode is much longer than the others. A normal episode of Anime is about 23
minutes, but the first episode of Fate Zero has nearly 48 minutes.
Through these two results, the rates of word and expressions in each level were different to each other. The rates
of words in each level were 35.2%(L4), 11.2%(L3), 22.0%(L2), and 11.2%(L1) in “Fate Zero.” The rates in “K-ON
2” were 46.9%(L4), 14.2%(L3), 18.7%(L2), and 5.2%(L1). The overall difficulty of K-ON is lower because the
conversations in K-ON are almost about relaxed daily dialogues.
The statistics also show the dialogues of Anime contained every level’s words and expressions. This
phenomenon may tell, even in one same episode, the degrees of difficulty of different scenes also could have large
differences. Therefore, we also need to separate the scenes and consider their own degrees. If we would like to
utilize Anime in supporting JFL learners who train their speaking and listening skills in the future, the estimation of
scenes' degrees also would be important.
Second, as we mentioned before, we collected the previous JLPT listening tests’ scripts as the standard
comparison data. We did the same statistics of the words in these speaking dialogues. Table 9 shows the number of
lines each year of previous JLPT listening tests’ scripts that we have collected.
The difference of total line numbers between different levels is mainly because the duration of listening tests is
different in each level. The L1 is the most difficult degree so that it includes the longest listening test. L1 is about 45
minutes. L2 is 40 minutes, L3 is 35 minutes and L4 is just about 25 minutes.
We used these listening tests’ scripts to make our statistics of Japanese words that appeared in each level. Table
10 shows the averages of words in the previous listening tests each level.
Through these results, we found that the rates of words of each level were different in each level’s test. In L4
tests, more than 94% words’ level was L4. L4 is the most primary level. The rates of words of L4 level were 79.55%
in L3 tests, 66.74% in L2 tests, and 59.27% in L1 tests. The rates decreased according to the proceeding of levels.
As the rates of L4 words decreased, rates of other levels’ words increased. The rates of L3 words ranged from
1.41% (in L4 tests) to 13.64% (in L1 tests). The rates of L1 and the rates of L2 words also changed according to the
proceeding of levels. This could give us an effective reference in estimating the dialogues’ degree of Japanese
Anime conversations. The rates of each word level of Anime dialogue were not similar with the rates of each word
level of the scripts of listening tests. We will compare the rates of words in subdivided dialogues with the scripts of
listening tests.
1352 Shan Junjie et al. / Procedia Computer Science 112 (2017) 1345–1352
Jungie Shan et al / Procedia Computer Science 00 (2017) 000–000

On the other hand, in the JLPT listening tests, there are also some types of conversation scenes in the dialogue.
Such as asking a question, explaining something and just making greetings to other peoples are appeared. When JFL
learners fail and have weakness in some types of practical dialogue scenes, they can make specialized trainings
through all the same types of scenes appeared in the dialogues of Anime. The classification for each Anime scenes
also would be a challenge.
In fact, the word lists and expressions lists were formulated by JLPT tests are just for the standard Japanese
language study. There are also lots of oral-flexible expressions and usages in the daily conversations of Anime
dialogues. To distinguish them from the Anime scenes or just judge them from the dialogues also could help the JFL
learners to understand the Japanese conversations more deeply. For the same reason, although we have matched the
words and expressions in exact JLPT levels, the degrees of conversations content would not be easily confirmed, not
like the reading contents or standard Japanese sentences or sections. The difficulties for them would be more fixable.
That’s also why we need to consider the past tests’ content as comparison data.

5. Conclusion

This paper reported the analysis results of words and expressions included dialogues of Anime and scripts of
listening tests of JLPT. We made subtitle texts of Anime and counted up the numbers of words and expressions for
each level. There were four level lists of the words, and there were three level lists of the expressions. We chose
three Animes with different genres respectively. We have found that the genres of Anime had effects on used words
and expressions. We made the scripts of listening tests of JLPT for five years. From the analysis results of the scripts,
we found that the rates of high-level words and expressions increased as the test’s level goes up. The rates of
included words and expressions for each level in the scripts were different from the rates in the dialogues of Anime.
We consider that the dialogues of Anime should be subdivided in order to match with the JLPT tests more exactly
and clearly. Then, the subdivided dialogues’ degree of difficulty is estimated.
For now, we have done the first and second step, and have made an analysis programs in doing words and
expressions’ statistics. We could use this method to analyze any new dialogue content of Anime and to make
statistics for JLPT tests’ texts. Next, we will try to subdivide Anime conversation scenes more clearly, and continue
to analyze the previous JLPT tests in detail and more deeply. Then use these statistic results to find out a standard
and effective measurement to estimate the dialogues’ degrees of Anime.

References

1. JAPAN FOUNDATION, Survey Report on Japanese-Language Education Abroad 2015, November, 10, 2016, URL:
https://www.jpf.go.jp/j/about/press/2016/dl/2016-057-1.pdf
2. Matsumoto H. Peak learning experiences and language learning: A study of American learners of Japanese. Language, Culture and Curriculum,
20(3): 195-208, 2007.
3. Kumano N, Hirokaga M. An Examination of "ANIME & MANGA": Regional Information Overseas and Japanese-Language Materials.
JAPAN FOUNDATION Japanese Education Bulletin, 2008 (4): 55-69 (in Japanese).
4. Manion A. Discovering Japan: Anime and Learning Japanese Culture. Unpublished master’s thesis. East Asian Studies Center, University of
Southern California, Los Angeles, CA. URL:http://www.chanpon.org/archive/manionthesis.pdf,2005.
5. Yazaki M. Development of "Anime-de Nihongo" Japanese Learning Activities with Anime: Examples from the Classes for Young JAL
Learners. Bulletin of the Shizuoka University International Center, 5: 57-74, 2011 (in Japanese).
6. Junjie S, Nishihara Y. Analysis of Japanese Animation Dialogue Level using Japanese Expressions. The Fifth Asian Conference on
Information Systems, Oct, 2016.
7. Keiko Kawashima and Nanae Kumano, Making Use of Anime and Manga for Japanese Language Classes, Nihongo Kyoiku Jissen Kenkyu
Forum Reports, 2011 (in Japanese).
8. Kara Lenore Williams, The Impact of Popular Culture Fandam on Perceptions of Japanese Language and Culture Learning: The Case of
Student Anime Fans, Doctor Thesis of the University of Texas at Austin, 2006.
9. Hazelbeck G, Saito H. A Corpus-based e-learning system for Japanese vocabulary. Information and Media Technologies, 2009, 4(4): 1104-
1128.
10. Li M, Ogata H, Hashimoto S, Adaptive Kanji learning using mobile-based email. Proceedings of the 17th international conference on
computers in education. 2009: 520.
11. Zhe L, Zhi S, Japanese Language Learning Assistant System of Constructing the Relation Diagram between Sentence Pattern and Vocabulary
for Chinese Learners of Japanese[J]. Japanese Society for Information and Systems in Education, 2015, 32(4): 246-256.

You might also like