Librenjak, Sara ORCID: https://orcid.org/0000-0003-1696-6777 (2022) Optimising kanji order in teaching Japanese using corpora. In: 15th Teaching and Language Corpora (TaLC) Conference, 13th – 16th July 2022, University of Limerick, Ireland.

Slideshow TALC 2022 SLibrenjak presentation.pptx - Accepted Version Download (8MB)

Abstract

This paper presents the research in the field of second language acquisition - Japanese language
teaching - using digital resources in order to optimise the learning order of the Japanese kanji script. The
Japanese language uses two syllabic scripts, hiragana and katakana, as well as a logographic writing
system called kanji, which consists of more than 2000 characters. Kanji script is considered one of the
most challenging areas in the acquisition of Japanese as a second language, and there is a plethora of
research on kanji learning methodology, strategies, as well as digital resources. However, there is
comparably little research on kanji learning order or criteria for including certain characters in textbooks
or curricula. Yan et al. (2013) and Loach and Wang (2016) present their algorithms for optimisation of
Chinese hanzi characters using a network approach and topological sort, but Japanese kanji, although
similar in shape, are more complex for foreign learners. Chinese hanzi characters usually have only one
reading, and there are no other scripts to combine with, whereas Japanese kanji usually have multiple
readings, and in combination with syllabic scripts this leads to a higher complexity and a huge number of
possible pronunciations.
Paxton (2015) and Kandrac (2021) tackle the topic of kanji learning order, but in this paper, we call for a
systematisation and corpus approach to the kanji ordering problem. In this research, will first discuss
various Japanese language corpora and the data about kanji frequency. We note that the frequency
changes according to the time span and genre of the texts in a corpus, so we argue that depending on
the learning objectives of the students, kanji order and choice should be different. Next, we will analyse
other factors that influence learning order besides the frequency in corpora. Third, we will propose an
algorithm to sort kanji based on learning objectives, their frequency in corpora, and their typological
properties. As this research aims to help Japanese teachers and learners by systematising a complex
issue of ordering kanji for optimal acquisition, we will lastly present the results of the experiment
conducted with university students of Japanese Language and Culture (N=43) where they compared the
order optimised using the algorithm, and the order from the textbook they have been using, Minna no
Nihongo 1. We found that after two semesters of study, the majority of students preferred the
reordered list. Thus, we conclude that the learning order of kanji can be improved using an algorithm
that takes into account corpus frequency and typological data about characters and that it may lead to
greater satisfaction and success in students.

Item Type:	Conference or Workshop Item (Paper)
Status:	Published
Subjects:	L Education > L Education (General) P Language and Literature > PL Languages and literatures of Eastern Asia, Africa, Oceania > PL501-699 Japanese language
School/Department:	School of Education, Language and Psychology
URI:	https://ray.yorksj.ac.uk/id/eprint/9829

University Staff: Request a correction | RaY Editors: Update this record