l2 learners' attitudes toward, and use of, mnemonic strategies when learning japanese kanji

L2 Learners’ Attitudes Toward, andUse of, Mnemonic Strategies WhenLearning Japanese KanjiHEATH ROSETrinity College, The University of DublinSchool of Linguistics, Speech and Communication SciencesCentre for Language and Communication StudiesRoom 124, 7–9 South Leinster StreetDublin 2, IrelandEmail: [email protected]

This study investigated kanji learning (the memorization of Japanese written characters) of universitystudents of Japanese, in order to evaluate students’ use of mnemonic strategies. The study applied in-depth qualitative methods to broaden the understanding of how foreign language learners usemnemonics when learning kanji. Data were collected over the duration of a year in the formof interviews,stimulated recall sessions, and a questionnaire on mnemonic usage. The study found that whilemnemonics are useful to memorize kanji and kanji components when applied in a meaningful way, anoverreliance on this strategy can have negative effects for the learner. The study highlighted numerousaccounts of themeaning of a kanji being lost in overly complexmnemonic strategies. Another limitationof mnemonic strategies was associations being made with the meaning of kanji and not with how it wasread, causing an inability to read kanji in Japanese.

Keywords: kanji; Japanese; mnemonic strategies; cognition; SLA

RESEARCHERS WIDELY AGREE THAT FOR-eign language learners from an English-speakingbackground have greater difficulty learning Japa-nese compared to learning European languages(Bourke, 1996; Douglas, 1992; Dwyer, 1997;Everson, 2011; Sayeg, 1996; Toyoda, 1998, 2000;Toyoda & Kubota, 2001). One study found thatuniversity students of Japanese progress moreslowly in language development than students ofother languages in terms of overall proficiency(Dwyer, 1997). In fact, Walton (1993) claimedthat it takes English-speaking students of Japanesethree times as long to acquire the same level ofproficiency as European languages. More recent-ly, Everson (2011) used a study of language

training required for State Department employ-ees (Jackson &Malone, 2010) to suggest Japanesetakes at least four times as long to acquire asEuropean languages do. Another study has linkedstudents’ difficulties in learning Japanese to ahigh attrition rate in Japanese language programsat Australian universities—especially in the case oflearners from an English-speaking background(Kato, 2002). Dwyer (1997) also found students ofJapanese were not reading as much as students ofEuropean languages, due to the challenges of theJapanese writing system. Indeed, Everson (2011)claims that the character-based writing system“presents special challenges for learners whosefirst language (L1) employs the Roman alphabet”(p. 251).

Kanji are logographic characters that originat-ed in China and represent language throughmorphemes (meaning-based units) rather thanphonemes (sound-based units). Due to thehistorical adaptation of kanji to the Japanese

The Modern Language Journal, 97, 4, (2013)DOI: 10.1111/j.1540-4781.2013.12040.x0026-7902/13/981–992 $1.50/0© 2013 The Modern Language Journal

language, most kanji have developed multiplereadings, which are defined as ways to pronouncea single character depending on its context anduse, thereby adding an extra challenge for theJapanese language learner. There are more than10,000 kanji in written Japanese; at the same time,knowledge of the 2,000most frequently occurringkanji enables language users to attain functionalliteracy in the language, because they make upmore than 99% of kanji in printed documents(Cook & Bassetti, 2005).

The difficulties of learning kanji have been welldocumented, especially for learners from analphabetic language background. Much researchinto kanji learning, therefore, has exploredmethods or strategies that can assist learners tomemorize kanji more effectively and efficiently.One aspect of memorization that has interestedresearchers is the examination of mnemonicstrategies. From a cognitive perspective a mne-monic strategy is defined as “a method forenhancing memory performance by giving thematerial to be remembered a meaningful inter-pretation” (Anderson, 2005, p. 461).

Since Japanese language education began toboom in the early 1980s, the commercial textbookmarket has seen a number of books that promotemnemonic strategies as a way to help studentslearn kanji more effectively (e.g., Heisig, 2007,2008; Henshall, 1988; Rowley, 1992; Stout &Hakone, 2011). Rowley’s book encourages the useof mnemonics with a pictorial strategy, whichinvolves relating the kanji or kanji components topictures of what they represent.While this strategyseems easy to apply to kanji that are symbolicrepresentations of their meaning, matters are lessstraightforward with more complex or abstractkanji, as is evident from Rowley’s (1992) ownadmission: “Several common kanji have beenexcluded [from the book] because frankly Icouldn’t come up with a satisfactory visual ortextual mnemonic” (p. 8).

Heisig’s (2007) book—now in its 6th edition—lays out amore systematic approach tomnemonicuse by breaking kanji into components thatinclude what he terms “primitive elements” (p.13), defined as the reoccurring components thatare present in a number of kanji that may giveclues to the kanji’s meaning. In studying kanjilearning, researchers favour the term grapheme(Toyoda, 2000) or component (Flaherty & Nogu-chi, 1998) to describe the smallest meaningfulunit that a kanji can be broken into. Althoughboth terms are used interchangeably, this articleuses the term component. It is important to notethat a component may or may not be the radical

(root component of a kanji for classificationpurposes). For example, a learner could break thekanji [language] into the components[speak], [mouth] and [five], of which only

[speak] is the radical. In Heisig’s (2007) bookthe kanji are presented with stories connectingthese components to hint at the kanji’s meaning,although little reference to the kanji’s reading inJapanese is made. That is, while a mnemonicstrategy may be used to connect the kanji[below] to its meaning, the same mnemonicstrategy provides no indication whether the kanjishould be read ge, shita, kuda, ka, or sa, all ofwhich are possible. Other authors, such as Stoutand Hakone (2011), go to great length to applycomplicated mnemonic strategies to even simplekanji. As an example, the authors teach the kanjifor the number 9, as a pictographic representa-tion of “a hand reaching something—nine is thenumber before reaching ten” (Stout & Hakone2011, p. 18). Here, the learners are encouraged touse a mnemonic strategy in conjunction with apictorial strategy (associating kanji with images).At the more intermediate or advanced level ofkanji learning, the learner is often encouraged touse a mnemonic strategy with a componentanalysis strategy. A simple illustration of thisstrategy using the previously mentioned kanji forlanguage [ ] is the mnemonic ‘My mouth canspeak five languages’: It associates the kanjicomponents of speak [ ], five [ ] and mouth[ ] with the meaning of language. Heisig (2007)argues that students might be embarrassed toapply a mnemonic strategy due to its “academicsilliness” (p. 3), but an organised approach tousing mnemonic strategies can help studentsattain the goal of “native proficiency in writing theJapanese characters” (p. 7).

Despite these authors’ assertion of the benefitsof mnemonics in kanji learning, none of the citedbooks provides empirical evidence that theirtechniques work, nor do they make reference toacademic studies that have explored mnemonicstrategy use in the learning of kanji. In one suchstudy, Lu et al. (1999) found that learnerslearned kanji more effectively when the kanjiwere introduced with mnemonic strategies thanwhen kanji were presented without such mne-monics. This quasi-experimental study highlight-ed the effectiveness of usingmnemonic strategies,but its results are not easily generalizable to reallearning environments where students oftenemploy multiple strategies at the same time. Intwo Australian studies of university students,Toyoda (1998) and Burke (1996) both foundmnemonics to be useful to students. Indeed, the

982 The Modern Language Journal 97 (2013)

learners in Toyoda’s (1998) study ranked mne-monics as the secondmost used of four strategies.However, that finding must be interpreted withsome caution inasmuch as the highest ranked ofthe four strategies was the analysis of a kanji’scomponents, which is a strategy often usedalongside, and not in opposition to, mnemonicstrategies. Nevertheless, these three studies illus-trate the benefits to memorization that mnemon-ic strategies have for the kanji learner.

Less encouraging of the benefits of mnemonicsare two studies that detected little significantimpact on a student’s ability to memoriselogographic scripts. Sakai’s (2004) study exam-ined the learning of kanji through mnemonics,compared to learning through context and rotememorisation; Wang and Thomas’s (1992) studycompared mnemonic strategies to rote learning.Once more results must be interpreted withcaution, because these two strategies are notmutually exclusive and can be employed at thesame time. Specifically, from a cognitive perspec-tive, mnemonic strategies are used to enhanceencoding of new information, and rote learningretrieves this encoded information. These ap-proaches, therefore, are looking at two verydifferent parts of the cognitive process. Becauseprevious research has yielded mixed results, notleast for methodological reasons, it is worthinvestigating mnemonic strategies from a some-what different vantage point in order to come tounderstand them more fully.

METHODOLOGY

Research Method

The research being reported here used amixedmethod qualitative approach in a case studycontext. In that decision it responds to bothcritics and defendants of language learningstrategy research who emphasize the importanceof qualitative research that goes beyond thefrequent reliance on questionnaires. Thus, ac-cording to Tseng, Dornyei, and Schmitt (2006),“researchers need to apply other, more qualitativemethodologies (such as stimulated recall andstructured observation) to achieve a fuller under-standing of the whole picture” (p. 98). Strategyuse is not only a highly individualized action, it isalso environment- and context-specific (Macaro&Grenfell, 2007; Rose, 2012b; Rubin et al., 2007;Woodrow, 2005) and therefore requires a highlyindividualized approach. This justifies the meth-odological choice of a case study approach, whichfacilitated the kind of in-depth and contextual-

ized exploration of mnemonic strategy use thathas largely eluded previous studies.

Instrumentation

For data collection the study used semistruc-tured interviews, stimulated kanji recall tasks, anda simple questionnaire; the questionnaire was notsubjected to statistical analysis but was used totriangulate data and as stimulus for discussion ininitial interview (see the subsequent discussion).The study itself is part of a larger research project(see Rose, 2010) that investigated a range ofaspects connected to kanji learning includingcognitive processes, the self-regulation of kanjilearning, motivation, and affect (see Rose, 2012a;2012b, for articles on theoretical findings, andRose & Harbon, 2013, for results of the self-regulation aspect of the larger project). Bycontrast, the current article focuses on evidenceregarding mnemonic strategy use obtained fromthe three previously mentioned instruments.

Ten interviews were held over the course of anacademic year; these coincided with 10 kanji testsfor the participants over that time period.Participants were asked to elaborate on theirapproaches to learning kanji in advance of eachkanji test (see Appendix A for a full list ofinterview questions and follow-up prompts).Comments relating to mnemonic usage werecoded using NVivo data analysis software.

The inclusion of a stimulated recall task reflectsrecent recommendations such as those by Tsenget al. (2006), who observed “We must note at thispoint that increased utilization of stimulatedrecall methodology (cf. Gass & Mackey, 2000)offers a promising future research direction inthis area, but so far little research has been donein this vein” (p. 82). Mackey and Gass (2005)define stimulated recall as “an introspectivetechnique for gathering data that can yieldinsights into a learner’s thought processes duringlanguage learning experiences. Learners areasked to introspect while viewing or hearingstimulus to prompt their recollections” (p. 226).For this study, the stimuli used in the recall taskwere the kanji tests based on kanji lists thestudents had studied in the preceding weeks.These tests required the students to write themeaning in English and reading in Japanese ofeach kanji, which appeared in the context of asentence. Participants were asked to describe theprocess by which they could recall each kanji thatappeared on the test. In other words, the specificprocess involved retrieval, which Matlin (2005)

Heath Rose 983

defines as locating and accessing information inmemory: Students access each kanji via theassociations used to encode the kanji whenlearning, thus giving clues on how the kanji hasbeenmemorized. Because written production wasnot required on the tests, it was excluded from thestudy and its design considerations.

Similarly differentiatedmethodological consid-erations informed the inclusion of a question-naire, on which students reported their strategyuse. Items on the questionnaire were adaptedfrom Bourke’s (1996) Strategic Inventory ofLearning Kanji in a fashion that required theprovision of answers on a linear scale, as opposedto being frequency-based. That decision reflectedrecent fundamental critiques of strategy invento-ries such as Oxford’s (1990) SILL by researcherslike Dornyei (2005) and Woodrow (2005). Asalready mentioned, no statistical analysis wasconducted on the questionnaire data, therebyavoiding psychometrically compromised ap-proaches to research (Dornyei, 2005). Instead,participant responses on the questionnaire pro-vided useful stimuli during the initial interview toelicit in-depth answers from participants regard-ing their strategy use.

Setting

The study was conducted in a university in thegreater Tokyo area over the course of an academicyear. That setting provided two beneficial quali-ties for case studies research. First, the universityreceived a large number of foreign students onyear-long exchanges, thus allowing the study toselect from a large pool of students who werestudying Japanese in a similar context; second, theprogram offered kanji courses separate from theremaining Japanese language curriculum, thusisolating the kanji learning task from otherlanguage skills.

The study excluded students from a languagebackground that used a logographic script, such asChinese, and participants needed to have a highcommand of English in order to take part in theinterviews, which were conducted only in English.

Participants had four 90-minute lessons perweek of Japanese language instruction, but kanjilearning was assigned as an autonomous learningtask. That programmatic decision was motivatedby the fact that proficiency in Japanese is notrelated to kanji knowledge. In other words, it ispossible to be a highly proficient speaker ofJapanese who is completely illiterate and viceversa. By making kanji learning autonomous, the

program allowed students to study at a level thatwas appropriate for them and to build up theirknowledge from that point forward. From aresearch perspective, such an arrangement de-creased the influence of teachers’ instruction onthe learning strategies the students employed.

As part of the kanji program, students weregiven lists of kanji to coincide with each of the 10kanji tests offered in the academic year. Althoughthese tests were conducted at the same time for allstudents, the kanji that appeared on the test wasdependent on the level of the kanji list that thestudents were studying.

As a researcher I was not involved in theplanning of the Japanese classes, had no part inconstructing the kanji lists or tests, and had norelationship with the participants beforehand.

Participants

Because previous studies have suggested arelationship between language proficiency andstrategy use (Bourke, 1996; Oxford, 2001), andbecause research into kanji learning has sug-gested that it is in the beginning stages of kanjilearning that students encounter most difficulty(Bourke, 1996; Sayeg, 1996), the study includedcomplete beginners as well as more advancedlearners.

A second consideration for choosing partic-ipants was the level of their perceived mnemonicstrategy use during the initial interview. At theoutset of the study I conducted a stimulated recalltask and initial interview with 23 potentialparticipants. Information gathered through thatactivity enabled me to select 12 on the basis of thelikelihood that they would contribute a broadrange of attitudes toward mnemonic strategy use.That decision reflects the observation by GrenfellandMacaro (2007) that “it is theoretically possibleto be a ‘good’ beginner language learner and a‘poor’ advanced learner” (p. 15). Because I wasinterested in a broad range of approaches Ipurposely included what might be seen asextreme or deviant cases. By that I mean studentswho would show extraordinarily high levels ofusage of mnemonic strategies or, in reverse,students who appeared to avoid mnemonicstrategies in kanji study altogether. By settlingon twelve participants the study’s precision,validity, and stability of findings could bestrengthened inasmuch as it would permitexamination of similar cases (Yin, 2002). At thesame time, it would avoid too “unwieldy a numberof participants for a single researcher to be able totreat the study with high complexity” (Miles &


Huberman, 1994, p. 30). Table 1 presents thetwelve cases, with pseudonyms replacing actualparticipant names.

RESULTS

Mnemonic Usage in the Stimulated Recall Session

Stimulated recall data revealed that mnemonicstrategy use among participants varied fromextreme frequency (Jacob and Simon) to con-

scious refusal to use them (Adrian), with theremaining cases falling between these two ex-tremes. The frequency of mnemonic strategyusage in the stimulated recall sessions is shown inTable 2, with the caveat that “qualitative analysis isnot about mere counting or providing numericsummaries” (Rubin & Rubin, 2005, p. 202).Because the students in each test were studyingdifferent lists of kanji, statistical comparisons onthese frequencies were not made. However, thesefrequencies provided insights and highlightedvarious patterns and tendencies of students’ useof mnemonic strategies.

The qualitative data from the stimulated recallsessions gave a richer understanding of howmnemonic strategies were used by students. Inparticular, mnemonics were used in conjunctionwith other strategies such as pictorial associationor, more commonly, component analysis. Anexample from a stimulated recall session of amnemonic strategy used with a pictorial associa-tion follows.

EXAMPLE 1

The verb taberu [eat]. Yeah. It looks like a house. So,for me, I eat in a house, or I eat in a restaurant, whichis like a building. So I would always think of it as thebuilding.

(Lisa, T4)

Here, Lisa associates the kanji [eat] as lookinglike a building using pictorial association. Shethen relates the meaning through a mnemonic of‘eating in a restaurant,’ which is a type of building.

Although pictorial associations were frequent,mnemonic strategies were often used with acomponent analysis strategy. That is, the

TABLE 1Participants

Proficiencyin kanji Participant

Country ofbackground

High Jacobþ UKSia UKHarold AustraliaAdrian� USA

Medium Craigþ USAPeter USAJerry USATim� Australia

Low Simonþ USACatherine USALisa UKSarah� USA

Note. Proficiency was based onnumber of known kanjiaccording to a placement test administered by theuniversity where the participants studied. Thisproficiency test was used to place students inappropriate level classes. “Low” indicates knowledgeof fewer than 200 kanji. “Medium” indicates knowl-edge of 200–800 kanji. “High” indicates knowledge ofmore than800 kanji.þ Indicates a participant chosenas a potential extreme case. �Indicates a participantchosen as a potential deviant case.

TABLE 2Frequency of Mnemonic Strategy Use in Stimulated Recall Sessions for Each of the 10 Kanji Tests

Name T1 T2 T3 T4 T5 T6 T7 T8 T9 T10 TOTAL

Jacob 20 18 20 19 20 20 20 19 20 20 196Simon 14 13 11 8 8 6 12 14 15 15 116Harold 8 6 11 4 7 4 7 7 3 5 62Sia 4 7 3 7 6 5 3 5 4 6 50Lisa 4 3 2 4 3 5 4 5 5 5 40Tim 4 3 5 2 4 3 4 2 4 1 32Sarah 0 4 3 2 4 2 0 1 3 2 21Craig 2 1 0 1 3 2 4 2 1 0 16Jerry 0 1 2 1 1 4 2 0 1 1 13Catherine 1 0 3 2 1 1 1 2 0 0 11Peter 2 1 0 2 0 0 0 1 2 3 11Adrian 0 0 0 0 0 0 0 0 0 0 0

Heath Rose 985

mnemonic strategy was connected with thecomponents of the kanji, as illustrated inExamples 2 and 3.

EXAMPLE 2

Like I always remember this one as yasumu [rest]because there’s a person resting under a tree. And it’snot really a story, but kind of—I know this is person,and I know this is tree. And I just try to picture theperson under the tree, and know that that’s relaxing . . .Or like suki [to like]. I know that that’s a mother and achild, and I say, what can be more pleasing than amother with her child? Like things like that. I don’t—Idon’t make up stories so much as I try to read thepictograph.

(Craig, T3)

EXAMPLE 3

That’s profit, so cutting down the stalk of rice forprofit.

(Simon, T6)

In these examples, Craig connects the kanji foryasumu [ ] to itsmeaning of rest bymaking a storyof its components of tree [ ] and person [ ] bythinking of a person resting under a tree [ þ

¼ ].Hegives another example of the kanji forlike, having the components of a mother [ ] andchild [ ], meaning like [ ]. Having studied theetymology (origin of the written form) of the kanjicomponents, Simon uses a mnemonic to associatethe etymologically based component of rice torecall themeaning of profit, in the act of harvestingrice for profit. The stimulated recall data was richwith examples of mnemonic strategies reminis-cent of similar examples described in previousstudies (Bourke, 1996; Lu et al., 1999; Rose, 2003).

Beyond revealing such strategy use, the stimu-lated recall data showed that mnemonic strategyusage occurred at all levels of proficiency. Forexample, beginner learner Simon and advancedlearner Jacob both showed high use of mnemonicstrategies (see Table 2). Similarly, infrequentusers of mnemonic strategies (e.g., Adrian,Catherine, and Peter) came from all proficiencylevels (High, Low, and Medium, respectively).

In light of these potentially unexpected find-ings, the interview data add important insights tothe study.

Resistance to Viewing Mnemonics as ‘Stories’

In the interviews, some participants made thedistinction between meaningful mnemonics (re-lating the meaning to actual components thathave meaning) and ‘stories’ that are less meaning-

ful. In fact, the stimulated recall data had very fewinstances where participants employed a mne-monic strategy that was not also based on anassociation ofmeaningful kanji components. Withthe exception of Jacob (who will be discussed insome depth later), participants developed aconsensus that mnemonics were useful onlywhen meaningful; otherwise they presented ob-stacles to language learning.

Many of the participants were keenly aware ofthe use of mnemonics as a strategy in kanjilearning even as they expressed strong opinionsregarding the limitations to studying kanji in thisway. Lisa, for example, understood the worth ofmnemonics but recognized the difficulty ofmaking meaningful connections, and their limi-tations for writing.

EXAMPLE 4

Yeah. That’s [using mnemonics is] the only way tomake it more interesting, and for it to make sense.Although what I find difficult is making up stories.And it’s useful for reading kanji, but when you’retrying to write it down, it’s a bit hard to recall the story.So—hmm.

(Lisa, T2)

On a slightly different note, Simon, who usedmnemonics regularly with his autonomous studyof kanji etymology, did not want to label hislearning as using ‘stories’ because of the stigmathat using stories is “confabulated”.

EXAMPLE 5

Interviewer: Okay. Ah, so this is what you’ve said withthe stories.

Simon: Stories I don’t do. It’s—that’s just not atactic I have tried or thought of trying. It’snot one I think would be very helpful forme, either.

Interviewer: Right. Right. There is—in a way, theetymology stories are kinds of stories.Right?

Simon: Yes. But I’d classify that differently. Youknow, I was thinking more along the linesof confabulation or something like that.

Interviewer: Okay.Simon: Stories is like here’s a kanji, and here’s the

picture that it’s showing, and here’s what’sgoing on in that picture, even though thatmay or may not have anything to do withthe etymology of the kanji.

Interviewer: Mm-hmm. Okay.Simon: Whereas, you know, I’m—I’m trying

to remember facts about the kanji asopposed to invent a story and thenremember that.

(Simon, T6)


It is then not surprising that in his self-reportquestionnaire (see Appendix B), Simon hadgiven his ownmnemonic strategy usage the lowestrating on the scale. In reality, however, Simon’suse of stories based on etymology, such as “cuttingthe rice to make a profit” were a mnemonicstrategy, which he used frequently (Table 2).

While Simon considered mnemonics to be“confabulated,” Peter characterized mnemonicsas potentially “ridiculous,” even as he acknowl-edged their usefulness for some learners in thememorization of kanji. This is illustrated with thefollowing example.

EXAMPLE 6

At first it sounds ridiculous, if someone explains it toyou: “The bunny comes out of the hole, goes around.The cat chases it through the grass to the rice field”.But it still makes sense. I can remember neko [ —cat] because of that [story].

(Peter, T1)

Nevertheless, despite stating the advantages ofmnemonics, Peter displayed very few uses ofmnemonic strategies in the 10 stimulated recallsessions (Table 2).

Yet more distanced toward the use of mnemon-ics was Adrian who expressed a complete dislike ofthem.

EXAMPLE 7

I don’t really try to use mnemonic devices oranything. Yeah. I’ve just never really tried. It justseems like you’re sort of going out of your way. I don’tknow. It just for me it’s not really natural. It just seemskind of time consuming. Because a lot of times themeaning doesn’t really fit in with—I don’t know. Itjust kind of depends. I’ve just kind of gone to justtrying to remember it. I don’t think I’m creativeenough to make stories for all of these.

(Adrian, T1)

This view was corroborated in questionnairedata, where Adrian rated his use of mnemonicstrategies at the lowest rating on the scale(Appendix B). However, unlike the case of Simonwho actually used them frequently, not one countof a mnemonic strategy was found in any of thestimulated recall sessions with Adrian (Table 2).This separated Adrian from other participantsbecause all other participants recorded multipleuses, despite voiced resistance in interviews.

The Danger of Overreliance on Mnemonics

At the other end of the spectrum of mnemonicstrategy use is Jacob who wholeheartedly embracedmnemonics. In stimulated recall sessions, Jacob

relied on mnemonic strategies for almost all kanjirecalled (see Table 2), a reflection of deliberateemployment of mnemonic strategies that he hadlearned in a book about kanji learning, calledRemembering the Kanji (Heisig, 2007). As previouslymentioned, it teaches mnemonics for kanjicomponents by relating them to their meaningin English and then using more mnemonicstrategies to relate these components to each otherin advanced or complex kanji. The results areindeed akin to the “confabulated” stories that theother participants (e.g., Lisa, Simon, and Adrian)seemed to be wary of using. Jacob describesHeisig’s approach in the following example.

EXAMPLE 8

So, sometimes he’ll stick very closely to what youmight find in the dictionary and sometimes he’ll giveit a meaning which it just—it’s not connected but it’smemorable. So we’ve got the kanji here. And as it says,“The picture in this kanji is not a pleasant one. Itshows a large and fluffy Saint Bernard dog stretchedout on the table all stuffed and stewed and garnishedwith vegetables—its paws in the air and an apple in itsmouth. At each corner of the table sits an eager butempty mouth waiting for the utensils to arrive so thefeast can begin.” Now, because that’s such a ridicu-lous story, it’s incredibly easy to remember. You’ve gotthese four mouths around the table, dog in themiddle, with their utensils. So that’s the kind of thingit’s based on. So the more ridiculous, the better forremembering it.

(Jacob, T3)

In the stimulated recall sessions, Jacob employedthis strategy to try tomemorize and recall all of thekanji in all of the sessions, using stories from thesame book to memorize kanji or their compo-nents and link these components together. In theinterviews, Jacob explained that this mnemonicstrategy filled most of his kanji study time. Forexample, he would often copy the stories onflashcards to test himself. He would also recordthe stories onto a digital audio player and listen tothem in bed or on the train, leading him to beable to memorize complicated and detailedstories such as the one in the excerpt above.

According to Jacob, the reason for his highreliance on mnemonic strategies was that sixmonths earlier he had become depressed anddisheartened with the kanji learning task and hadviewed the task of learning 2,000 kanji beforegraduation to be an unachievable goal. When afriend introduced him to the Heisig learningapproach, he found that mnemonics gave him asystematic way to study kanji, whichmotivated himto study, and made his goals seem achievable.

Heath Rose 987

At the beginning of the study Jacob was veryenthusiastic about learning kanji in this way. Butthat feeling diminished throughout the study sothat toward the end Jacob was once again feelingthat he could not attain his goals because hebegan to notice serious limitations in theapproach. The biggest limitation remained thatthe stories focused on English meanings of kanjipresented in isolation. In many cases, especially atthe advanced level, these kanji rarely appeared bythemselves as a reflection of their singularmeaning, but in common combinations withother kanji. Thus, often when known kanjiappeared in text, Jacob was able to recognizethe kanji, its story, and its singular (and at timesabstract) meaning; but this did not help him tounderstand the meaning of it in the context of asentence. Moreover, these stories, which werealways entirely in English, made no associationwith how the character was read in Japanese,resulting in an inability to read known kanji in theJapanese language, or to look them up in aphonetic dictionary. Although he hadmemorizedthemeaning of the kanji, he had not at all studiedthe readings of them and came to realize that heneeded to go back and relearn all the kanji andmake associations with their Japanese readings.This left him disheartened that the hard work hehad put into remembering the kanji had notresulted in the level of progress that he had hopedfor.

Losing the Meaning in the Mnemonic

Yet more problematic, Jacob could at timesrecall the mnemonic strategy, but not themeaning, as in the following example.

EXAMPLE 9

Some of them are quite ridiculous. “There’s thismonster on the ceiling that likes to eat nails and thenspit them in people’s heads.” I remember the story,but I can’t remember the kanji right now. I haven’trevised these [kanji] now for about three months orlonger. I find after three months it’s terrible.

(Jacob, T8)

The same phenomenon of losing the meaning inthe mnemonic was evident with other learners aswell, especially with frequent mnemonic strategyusers. Simon would often remember the etymo-logical origins of kanji components and thenforget the meaning of the kanji they werepresented in—especially if the meaning of the

kanji had grown apart from its etymologicalorigin. Catherine, who made her own mnemonicstrategies, would regularly create stories accord-ing to pictorial associations of a kanji’s compo-nents that were completely removed from themeaning, as in the following example.

EXAMPLE 10

Here, Catherine remembers the story, but itfails to connect with the meaning of the kanji inhermnemonic that is, her story of a tree in a box isunassociated with the kanji for komaru [ ],meaning to be in trouble. In another case, Catherinememorizes the kanji [feeling], because it “lookslike a bug that’s been squashed,” but was unableto recall its meaning. This was a reoccurringphenomenon in Catherine’s application of mne-monic strategies and differentiated her fromother participants who tried to make meaningfulassociations with mnemonics, even whenconfabulated.

In sum, evidence for the phenomenon of losingthe meaning in mnemonic arose a number oftimes in the stimulated recall sessions, especiallywith participants who relied heavily and con-sciously on mnemonic strategies. It tended to bethe result of making stories that were removedfrom the meaning of the kanji. That is, saying akanji looks like a monster on the ceiling spittingnails into people’s heads does not necessarilyevoke its real meaning, an insight to which someparticipants gave voice. In fact, Peter commentedon this during one of the initial interviews.

EXAMPLE 11

Here Peter, like Simon, emphasizes the use ofmnemonics only when it makes sense, and avoids

Catherine: Yeah. Because I specifically rememberthat one [ ] as a tree [ ] in a box[ ].

Interviewer: Okay. But how does that link in with themeaning of trouble? Or you just-

Catherine: I don’t know what the meaning is.

(Catherine, T9)

Interviewer: If there’s no clear connection betweenthe compounds, you wouldn’t use thistechnique?

Peter: No.Interviewer: At all?Peter: No. Because like maybe that helps in

memorizing them, but if I’m trying toread something, you know, you don’t havetime to sit there and go, okay, the bunnycame out of the hole, so that kanji mustmean this. You know?

(Peter, T1)


applying mnemonics in nonsensical or “confabu-lated” ways.

DISCUSSION AND CONCLUSION

Implications for Learners and Instructors

While mnemonic strategies were clearly usefulto a number of learners in this study, data revealedlimitations to the exclusive use of a mnemonicapproach. Such findings have clear implicationsfor the Japanese language learner and instructor.

Data indicated that mnemonic strategies wereuseful when applied in ameaningful way, but wereless helpful when associations became convolutedor complex. Results indicate convoluted mne-monic strategies, which link advanced kanji tomeaning andnot the Japanese reading,may resultin an inability for the learner to read the kanji inJapanese, or recall the meaning altogether.Results also indicate adherence to a systematicapproach of mnemonic strategies can lead tofeelings of defeat when knowledge of kanji doesnot advance as expected.

While this study warns against the shortcomingsof a mnemonic approach, it does concur withprevious studies that show mnemonics to be auseful strategy for students of Japanese (e.g., Luet al., 1999; Toyoda, 1998), but qualifies thisfinding by arguing that it is only useful whenmeaningful associations can be made, whichtrigger both the meaning and the reading ofthe kanji. Logically, mnemonics might prove lessuseful for kanji that are rarely used in isolation ofother kanji, or are semantically abstract, as is thecase formany advanced kanji. Therefore, teachersand learners alike should be wary of commercialtextbooks that claim a systematic approach tomnemonics for all kanji can help learners attainnative proficiency in written Japanese. Certainlysuch claims cannot be supported by this study orby the findings of previous empirical research intokanji learning strategies.

Implications for Further Research

This study began by identifying numerouscontradictory findings regarding the benefits ofmnemonic strategies in kanji learning and pro-ceeded to explore mnemonic strategy use from adifferent research approach than previous studies,which trended towards quasi-experimental orsurvey designs. By situating the examination ofmnemonics in a larger study of strategy use (Rose,2010), this article examined how learners usemnemonic strategies alongside other strategies.

Future studies might also consider an approachwhere use of mnemonic strategies can be mea-sured in conjunction with (and not in oppositionto) strategies such as component analysis andpictorial association.

In terms of data collection, stimulated recalldata shed light on the types of mnemonicstrategies employed by learners, concurring withrecent recommendations for the inclusion of thisinstrument in strategy research (Dornyei, 2005;Rose, 2012b; Tseng et al., 2006). Qualitativeinterview data revealed the impetus behindchoosing or avoiding mnemonic strategies, andthe self-perceived impact they had on learning. Bycomparing data from the questionnaire andinterview with the stimulated recall task, limita-tions to self-report instruments were also revealedin the form of learners who reported not usingmnemonic strategies but that were neverthelessobserved in stimulated recall and vice versa. Thisconcurs with previous studies that have warnedagainst the reliability of self-report data in strategyresearch (Rose, 2012b; Tseng et al., 2006; Wood-row, 2005).

A limitation of the study is that the contextinvolved the learning of kanji lists for readingtests, and thus the important role of readingauthentic texts in the learning process (Everson,2011) and the written production of kanji werenot included in the study design. Furthermore, asthe participants in this study were learningdifferent lists of kanji, analysis of mnemonicstrategy use according to kanji was not possibleacross cases, and thus a future study that examinesa cohort of learners engaged in the same level oflearning could reveal interesting results regardinga correlation between types of kanji and mne-monic strategies.

CONCLUSION

Results of the study have suggested thatmnemonic strategies provide both benefits andlimitations in the kanji learning task. In terms ofattitudes toward this cognitive strategy, whilesome participants perceived mnemonics as auseful strategy for kanji learning, others perceivedthem as “ridiculous” stories. A further examina-tion of the cases revealed that an overuse ofmnemonic strategies caused limitations in learnerrecall of kanji due to the multiplicity of kanjireadings. The findings not only support the use ofmeaningful mnemonic strategies, but also warnJapanese language learners and instructors of thedangers of relying too heavily on them. This studyalso exposes limitations to the sole use of

Heath Rose 989

mnemonic strategies, which are encouraged by anumber of commercial kanji learning textbooks.Finally, the study has highlighted implicationsfor future kanji research, and the findingssupport movements toward interviews and stimu-lated recall data collection instruments whenlooking at kanji learning, in place of the self-report questionnaires that have traditionallybeen used.

ACKNOWLEDGMENTS

I would like to express my sincere gratitude to theanonymous MLJ reviewers for their insightful com-ments, and toHeidi Byrnes for her valuable input on thisarticle.

REFERENCES

Anderson, J. R. (2005). Cognitive psychology and itsimplications. (6th ed.). New York: Worth.

Bourke, B. (1996). Maximising the kanji learning task.(Unpublished doctoral dissertation). The Univer-sity of Queensland, Brisbane, Australia.

Cook, V., & Bassetti, B. (2005). Second language writingsystems. Clevedon, UK: Multilingual Matters.

Dornyei, Z. (2005). The psychology of the language learner:Individual differences in second language acquisition.Mahwah, NJ: Lawrence Erlbaum.

Douglas, M. O. (1992). Development of orthography-relatedreading and writing strategies by learners of Japanese asa foreign language (Unpublished doctoral disserta-tion). The University of Southern California, LosAngeles, CA.

Dwyer, E. S. (1997). Getting started the right way: Aninvestigation into the introduction of kanji study toneophyte Japanese learners (Unpublished doctoraldissertation). The University of Texas, Austin,TX.

Everson, M. E. (2011). Best practices in teachinglogographic and non-Roman writing systems toL2 learners.Annual Review of Applied Linguistics, 31,249–274.

Flaherty, M., & Noguchi, M. S. (1998). Effectiveness ofdifferent writing approaches to kanji educationwith second language learners. JALT Journal, 20,60–78.

Heisig, J. W. (2007). Remembering the kanji, Vol. 1: Asystematic guide to reading Japanese characters. Hon-olulu, HI: University of Hawai’i Press.

Heisig, J. W. (2008). Remembering the kanji, Vol. 2: Asystematic guide to reading Japanese characters. Hon-olulu, HI: University of Hawai’i Press.

Henshall, K. G. (1988). A guide to remembering Japanesecharacters. North Clarendon, VT: Tuttle.

Jackson, F. H., & Malone, M. E. (2010). Building theforeign language capacity we need: Toward a compre-hensive strategy for a national language framework.Washington, DC: Center of Applied Linguistics.Retrieved 10 December 2012, from http://www.cal.org/resources/pubs/building-the-foreign-language-capacity-we-need.html

Kato, F. (2002). Efficacy of intervention strategies inlearning success rates. Foreign Language Annals, 35,61–72.

Lu,M.,Webb, J.M., Krus, D. J., & Fox, L. S. (1999).Usingorder analytic instructional hierarchies of mne-monics to facilitate learningChinese and Japanesekanji characters. The Journal of Experimental Educa-tion, 67, 293–311.

Macaro, E., & Grenfell, M. (2007). Claims and critiques.In A. D. Cohen & E. M. Macaro (Eds.), Languagelearner strategies (pp. 9–28). Oxford: OxfordUniversity Press.

Mackey, A., & Gass, S. (2005). Second language research:Methodology and design. Mahwah, NJ: LawrenceErlbaum.

Matlin, M. W. (2005). Cognition (6th ed.). Hoboken, NJ:John Wiley & Sons.

Miles, M. B., & Huberman, A. M. (1994).Qualitative dataanalysis. Thousand Oaks, CA: SAGE.

Oxford, R. L. (1990). Language learning strategies: Whatevery teacher should know. Boston: Heinle & Heinle.

Oxford, R. (2001). Language learning strategies. InD. Nunan & R. Carter (Eds.), The Cambridge guideto teaching English to speakers of other languages(pp. 166–172). Cambridge: Cambridge UniversityPress.

Rose, H. (2003). Teaching learning strategies forlearner success. Babel, 32, 32–38.

Rose, H. (2010). Kanji learning of Japanese languagelearners on a year-long study exchange program at aJapanese university: An investigation of strategy use,motivation control and self-regulation. (Unpublisheddoctoral dissertation). The University of Sydney,Sydney, Australia.

Rose, H. (2012a). Reconceptualizing strategic learningin the face of self-regulation: Throwing languagelearning strategies out with the bathwater. AppliedLinguistics, 33, 92–98.

Rose, H. (2012b). Language learning strategy research:Where do we go from here? Studies in Self-AccessLearning Journal, 3, 137–148.

Rose, H., & Harbon, L. (2013). Self-regulation insecond language learning: An investigation ofthe kanji-learning task. Foreign Language Annals,46, 96–107.

Rowley, M. (1992). Kanji pict-o-graphix: Over 1,000Japanese kanji and kana mnemonics. Berkeley, CA:Stone Bridge Press.

Rubin, J., Chamot, A. U., Harris, V., & Anderson, N. J.(2007). Intervening in the use of strategies. In A.D. Cohen & E. M. Macaro (Eds.), Language learnerstrategies (pp. 141–160). Oxford: Oxford Univer-sity Press.


Rubin, J. H., & Rubin, I. S. (2005). Qualitativeinterviewing: The art of hearing data. ThousandOaks, CA: SAGE.

Sakai, J. (2004). Imagery mnemonic versus context learning:Effects on kanji retention, (Unpublished mastersthesis). California State University, Long Beach,CA.

Sayeg, Y. (1996). The role of sound in reading kanji andkana: A review. Australian Review of AppliedLinguistics, 19, 139–151.

Stout, M., & Hakone, K. (2011). Basic Japanese kanjivolume 1: High-frequency kanji at your command.North Clarendon, VT: Tuttle.

Toyoda, E. (1998). Teaching kanji by focusing onlearners’ development of graphemic awareness.Australian Review of Applied Linguistics, 21, 155–168.

Toyoda, E. (2000). English speaking learners’ use ofcomponent information in processing unfamiliarkanji. Australian Review of Applied Linguistics, 23,1–14.

Toyoda, E., & Kubota, M. (2001). Learning strategiesemployed for learning words written in kanjiversus kana. Australian Review of Applied Linguistics,24, 1–16.

Tseng, W. T., Dornyei, Z., & Schmitt, N. (2006). A newapproach to assessing strategic learning: The caseof self-regulation in vocabulary acquisition.AppliedLinguistics, 27, 78–102.

Walton, A. R. (1993). Japanese language in US highschools: A new initiative. Modern Language Journal,77, 522–523.

Wang, A. Y., & Thomas, M. H. (1992). The effect ofimagery on the long-term retention of Chinesecharacters. Language Learning, 42, 359–376.

Woodrow, L. (2005). The challenge of measuringlanguage learning strategies. Foreign LanguageAnnals, 38, 90–100.

Yin, R. K. (2002). Case study research, design and methods.Thousand Oaks, CA: SAGE.

APPENDIX A: Interview Questions

The interview questions have been organizedinto main questions, which were always asked toall participants, and probes used to elicit moreinformation where necessary.Part One (For Initial Interview Only):

These questions concern students’ past expe-riences of studying Japanese.

i. Can I first ask you about your Japaneselearning background? Probes: How longhave you studied Japanese? / Where haveyou studied?

ii. What kinds of learning have you taken partin? Probes: What kind of Japanese classeshave you taken in the past? / How activewhere these classes? / What did a typicallesson consist of?

iii. How do feel about learning kanji in yourJapanese learning so far? Probes: Do youfind it difficult? Fun? Challenging? Over-whelming / How much do you alreadyknow? / How do you feel about studyingkanji in the future?

Part Two (for Initial and Ongoing Interviews):These questions concerned students’ kanji

learning in the preceding week(s).

i. Remember back to your last Japanese kanji/writing class. What did you do? Probes:How did you study them or learn them? /How were they introduced by the teacher? /Did the teacher show you ways to rememberany of the kanji more easily? If yes,what?

ii. How have you been studying kanji?Probes: How have you studied them? /How long did you spend in each studysession? / Probe for reflections on strategyuse, study habits, difficulties of learning,and external factors that may have affectedstudy, i.e., reasons for under or overperformance.

iii. Reflecting on your kanji learning in thepast week(s), how do you think you haveperformed?

iv. Can I finally ask you if there is any aspect ofyour experience of learning kanji that hasnot been covered in this interview?

Part Three (Stimulated Recall):This task examined the cognitive processes of

how students remembered and recalled kanjilearned in the preceding week(s).

i. I am going to show you a list of kanji,which you have been studying. I want youto tell me the reading or meaning ofthe kanji and any information on howyou remember it. Probes: How did youremember it? / How do you know that isthe answer? / Did you remember it in anyspecial way?

Heath Rose 991

APPENDIX B: Questionnaire Results

Q1 Q2 Q3

Adrian 1 1 1Craig 5 5 3Harold 2 5 1Jerry 7 2 1Jacob 3 7 7Catherine 4 2 4Lisa 4 4 4Peter 1 1 1Simon 1 1 1Sia 6 6 2Tim 3 5 6Sarah 6 3 2

Note: Q1 I make up my own stories according to whatthe kanji looks like to me; Q2 I make up my ownstories according to the component elements of thekanji;Q3 I use stories told to me by my teacher or inreference books; Scale 1¼not true of me at all;7¼ very true of me


l2 learners' attitudes toward, and use of, mnemonic strategies when learning japanese kanji

Documents