Hey, someone's paper refererences RevtK! ;p Is Lars Yencken one of ours (edit: Judging by past papers I doubt they used RTK but I swear the name's familiar, maybe I cited them once)?? RevtK now has academic cred.

The story used was something about smashing something 9 or 10 times with a stone.

http://lars.yencken.org/papers/phd-thesis.pdf

Edit: Searching Google Scholar for kanji.koohii.com turned up a few other results as well. Three papers by Norman Lin, Shoji Kajita, and Kenji Mase.

A Multi-Modal Mobile Device for Learning Japanese Kanji Characters through Mnemonic Stories

Abstract: We describe the design of a novel multi-modal, mobile
computer system to support foreign students in learning
Japanese kanji characters through creation of mnemonic sto-
ries. Our system treats complicated kanji shapes as hier-
archical compositions of smaller shapes (following Heisig,
1986) and allows hyperlink navigation to quickly follow
whole-part relationships. Visual display of kanji shape and
meaning are augmented with user-supplied mnemonic sto-
ries in audio form, thereby dividing the learning information
multi-modally into visual and audio modalities. A device-
naming scheme and color-coding allow for asynchronous
sharing of audio mnemonic stories among different users’
devices. We describe the design decisions for our mobile
multi-modal interface and present initial usability results
based on feedback from beginning kanji learners. Our com-
bination of mnemonic stories, audio and video modalities,
and mobile device provide a new and effective system for
computer-assisted kanji learning.

Mobile User Behavior and Attitudes during Story-based Kanji Learning

Abstract: We report on the usability and usage patterns of a mobile system designed to help students learn Japanese kanji characters through a mnemonic-story method . Our data comes from two sources: (1) a usability questionnaire of kanji learners, and (2) from a series of two-week experiments where subjects used our mobile kanji learning system . We found that 21% of mobile device usage actually occurred in a mobile learning setting . We also found varied time-of-day usage patterns over a two-week period . We discovered that user behaviors while using the system – selection operations, story creation time and story length – were characterized by positively-skewed distributions that can be described by exponential and lognormal models. We found statistically significant correlations between certain user behaviors and certain self-reported user attitudes. These results empirically clarify and confirm the usability of our mobile story-based kanji learning system . Questionnaire data from a separate population of story-based kanji learners, who did not use our device, provides further insight into the study behaviors and requirements of story-based kanji learners .

Collaborative story-based kanji learning using an augmented tabletop system

Abstract: We present a novel augmented tabletop com-
puter system to support collaborative story-
based learning of Japanese kanji characters by
non-native students of the Japanese language.
Our system is based on interactive use of physi-
cal and virtual media using Augmented Reality
technologies. To investigate the eff ectiveness
of this approach, we use Heisig’s decomposi-
tion of complex kanji into simpler component
parts and support constructionist learning
by allowing tangible exploration and physical
construction – analogous to construction of a
jigsaw puzzle – of complex kanji shapes from
simpler component shapes which are printed
on physical cards. Cards can be associated
with video content that is projected onto the
card’s location, thereby augmenting the physi-
cal kanji component with virtual media. A col-
laborative pen-based handwriting interface al-
lows students to create mnemonic stories, in
the form of pictures or words, to assist in mne-
monic memorization of complex kanji as col-
lections of simpler components. Experimental
evaluation of the system compared user learn-
ing behavior between experimental card-using
conditions and a control GUI GUI condition. For
card-using conditions, we observed increases
in exploratory activity and pointing behav-
ior; for the AR AR media condition, we observed
increased exclamatory clapping. Our system
demonstrates that story-based kanji learning
– which is typically an individual learning pro-
cess – can be realized as a collaborative, con-
structionist, and computer-assisted language
learning activity.  is points to the possibility
of story-based kanji learning technologies be-
ing used in future classroom scenarios.

Looks like an interesting paper, but it's a bit offputting that the first two pages start off with an exaggeratedly bleak representation of looking up kanji. He makes it sound like the only possible ways of looking up kanji are by pronunciation, or by using a paper dictionary to look it up by the main radical. Given that the paper was published this year, he must know there are other ways of looking up characters, right? In fact he mentions some of them in his paper, so I don't understand why he worded his hypothetical anectdote that way.

JimmySeal Wrote:Looks like an interesting paper, but it's a bit offputting that the first two pages start off with an exaggeratedly bleak representation of looking up kanji. He makes it sound like the only possible ways of looking up kanji are by pronunciation, or by using a paper dictionary to look it up by the main radical. Given that the paper was published this year, he must know there are other ways of looking up characters, right? In fact he mentions some of them in his paper, so I don't understand why he worded his hypothetical anectdote that way.

Yeah that's weird.

Quote:We finally combine kanji confusion and pronunciation models into a new adaptive testing platform, Kanji Tester, modelled after aspects of the Japanese Language proficiency Test. As the user tests themselves, the system adapts to their error patterns and uses this information to make future tests more difficult.

That sound pretty sadistic, actually. It also sounds a little bit like SRS, which aims to show you material you are most likely to have forgotten.

Quote:as graphemic input methods are not readily available for Japanese

Sure they are. I use one all the time, the "IME Pad" feature in Microsoft Japanese IME. I can't say I know of another one. However this program is available for free to an awful lot of people.

Quote:However, of the computer-based systems, many still suffer from several issues: the awkwardness of mouse input for drawing characters; sensitivity to both stroke order and connectivity of components; and the difference in hand-writing styles between learners and native speakers.

Like I said before I use MS Japanese IME and I speak English (so I'm obviously a learner). I never have trouble with the IME Pad feature. I also studied RTK1 so I know a little about stroke orders. IME Pad is only slightly sensitive to "number of strokes," and isn't sensitive at all to stroke order. I think it's only slightly sensitive to connectivity. I can get the right character every time, unless I draw it wrong.

Section 3.1 is entitled "The Learner's Paradox."

Quote:Coady (1997) calls this problem the “Beginner’s Paradox”, since learners must acquire vocabulary through reading, but need sufficient vocabulary to bootstrap the process.

I just thought that was a good quote.

Quote:Use of appropriate tools is also increasing; Tozcu and Coady (2004) found that teaching high-frequency words using a CALL system improved reading more than simple reading practice.

What is a CALL system?

...more thoughts later.

Quote:We discussed earlier the Heisig method as advocating a particular form of mnemonic, where a story is constructed for a kanji involving either aspects of its form or its component radicals. An innovative site supporting this study method is the Reviewing thee Kanji site. Heisig (1985) advocates learners developing their own mnemonic stories, however this can sometimes be difficult if the elements a learner must include in these stories seem unrelated.

Shout out!

jcdietz03 Wrote:What is a CALL system?

computer assisted language learning

Quote:Firstly, we know that word knowledge is multi-dimensional, yet the general-purpose nature of [Quizlet] flashcards limit users to studying a single dimension of knowledge at a time.

So, "flashcards suck" in so many words. Don't people say to make the amount you have to know for a single card as small as possible? Also, I remember reading about this somewhere else...

Quote:For all of these systems, the lack of a significant time-cost penalty for unknown words could have several undesirable consequences, in particular over-reliance on and abuse of the dictionary as a learning tool.

And he says earlier in the paper that dictionary lookup time is undesirable and should be minimized!

So they made a dictionary called FOKS. Actually a revision to a previous version of that same dictionary. They changed the way they did it from before, and considered similar characters this time instead of just wrong / incorrect readings. As explained previously, I would rather look things up the hard way. The #1 problem I have is "cannot read character from source material."

They made a test called Kanji Tester. And it is available here for you to use if you want: http://kanjitester.gakusha.info/
I haven't used it yet. The paper author says the system is bad for improving your vocabulary level and it is good for assessing your progress. For the JLPT syllabus, there is a large number of words. Each word can generate 1000s of different possible questions and question types, which will test various aspects of vocabulary knowledge. Therefore no two tests will be the same, and it is rare to see the same question asked twice because the syllabus is large and the questions and question types are random.

So what is the answer for how to escape the cycle? I never saw this kind of diagram before, but I think it is accurate.

This is the answer, but unfortunately lots of people don't have the patience for it at first. That's why things like graded readers exist, and can be useful.