Teaching English through English (TETE)
著者 Hayase Mitsuaki
journal or
publication title
Philologia
volume 40
page range 111‑126
year 2009‑02‑01
URL http://hdl.handle.net/10076/10596
Teaching English throughEnglish (TETE)
Mitsuaki Hayase
Imtrod uctioII
The present paper discusses several important issues related to TETE with the purpose of understanding what TETE is all about and finding what lEnglishteachers
&om JHS through UNIV in Japan are expected to do in order to put lifTeinto TETE.
ARer reviewlng the actualsituation in Japan, the paper reveals possible reasons behind the fact that TETE has not particularly permeated in this country, shows a lack of emplrlCal evidence of the efficacy of TETE, examines theories and hypotheses of TETE, and finally Elves Specific suggestions to carry out TETE. In this paper, TETE is operationally de血ed as "teaching English mostly tbrougb E喝1isb."
I. Actllal situation
How is English taughtinjuniorhigh schools (JHS),senior highschools (SHS),
and universities (UNIV) interms of 玉itheuse
ofEnglish? Do teachers teach their classes mostly in English or in Japanese? It is important toknow the actualsituation as tothe
use ofEnglish in English classes before a discussion of TETE begins.As for the actual situation at JHSand SHS, the Ministry of Education, Culture, Sports, Science and Teclmology (MEXT) has been conducting an annual survey on theuse of English by teachers and students since 2005, and the results will be shown in the next section pr10r tothe discussion on the actualsituation atUNTV.
1.2 Actual situation at JⅡS and SⅡS
(1)JES: Teachers
Table I Junior highschoolswhere more than half(1e丑side)/most(rightside)of
English class istaughtin English (Useof English by teachers inpercentage)
2004 2005 2006 2007
1st‑year 32ー0 3.9 31.5 3.8 32.2 3.6 30.9
2nd‑year 31.2 3.7 31.1 3.6 30.9 3.4 30.3
3rd‑year 30.0 4.0 29.7 3.9 29.3 4.2 29.2
Based on the MEXT (2005,2006, 2007,and 2008)
(2)SⅢS: teachers
Table 2 "Oral Communication I" and "English I" in senior high schools where more than half(le丘side)/most(dght side)ofEnglishclass jstaughtin Eng一ish (Useof Englishbyteacbersinpercentage
2004 2005 2006 2007
Oral 32.0 48.0 31.5 46.9 32.7 45.3 34.0
Commlmic血ionⅠ 31.3 25.1 33.3 21.5 31.8 21.5 33.9
EnglisbⅠ 20.3 3.7 23.1 9.9 25.7 9.6 24.3
7.7 1.1 8.4 1.1 8.3 1.3 10.0
Reading 14.6 3.7 16.1 6.0 15.6 2.7 13.l
+Upper row: Schools with International and English departments
*Lower row: All the other schools
Based on the M玉XT (2OO5,2006, 2007, and 2008)
(3)SHS: students
Table3UseofEn lishbystudentsinpercentage
2006 2007
Sometimes heveryJesson Sometimes heverylesson
Discussionanddebate ?8.1 4.3 27.7 5.5
in"Commmicabonr'
13.9 1.3 12.3 1.2
Presmting
infomlationandide8S 57.8 22.3 62.5 l9.8
in"CorrLmmiC&tiOnr' 51.3 ll.0 54.4 ll.1
Student‑studenttalk in"Commmicationr'
2S.8 68.3 29.2 66.4
43.6 44.0 40.9 49.1
Student‑studenttaJk 45.9 l8.8 48.2 17.7
in"EngJishr
*Upper row: Schools with lnte皿ational and English departments
*Lower row: All the other schools Based on the MEXT (2007,2008)
A close analysis of these tables reveals several intrlguing facts. First, both atJHS
∽d S班S, as tothe use orEnglisb by the teachers the overall percentages in the various classes did not change much throughout the suⅣeyed fわur years. In other words, there
was not a consplCuOuS increase or decrease in the use orEnglish by the teachers atJHS and SHS. Second, at JHS about 30% of the teachers conducted more than halfoftheir class in English even thoughonly about three percent conducted most of their classes in
English.Altogether,about one‑third of the teachers conducted atleastmore than half of their classes in English. Thoughone‑thirdisnot ahigh rate, it isa good startlng POlnt tO increase the percentage in the future. Third, at SHS, there isa big gap between schools with Intemational and English departments andall the other schools with the percentage of the former higher than that of the latter in all the subjectssurveyed. It is
understandable as those schoolswith Internationaland English departments are
generally more entlmsiastic and determined to use English as much as possible.
Fourth, as to "OralCommunication I" in schools with International and English
departments, about one third of the teachers used Englishin half of their classes while about halfof them used English inmost of their classes. Put together, about 80% of the teachers use English at least half of their classes, which createsanideal situation to increase the percentage of the teachers conducting most of the class inEnglishin the
future. Sixth, as to "English I" and "Reading", the percentages are not favorable at either of schools. Thepercentages in "English I" and =Reading" inal1 the other schools are particularly low; only about one or two percent of the teachers used Englishmost of their classes, afTarcry &om the proposal guideline issued by the MXET on D.ecember 22, 2008 (耳okono gakushushidoyoryo kaiteian)which stipulates that "all the English
classes at SHS should be conducted inEnglishas a pnnciple." If the NEXT is serious about the stlPulation, a drastic change in teaching methods, textbooks, examinations, and possibly teachers'skills in English and teaching methods is needed before thenew
course of study is put into use in 2013 with the first‑yearstudents. At present it is
known that compared to thecurrent textbooks, a血・ greater emphasis on communication skills will be put in thenew textbookswith the change of the names of textbooks in which the words like "communication", "expressions in English'',and "Erlglish
conversation" will be丘・equently used. Finally, as to theuse of English by students at SHS in 2006and 2007, by far the most common actlVlty in both kinds of schools is student‑toIStudent talk in "Communication I" shown in the percentage of about 40 to 70
(inevery lesson)and about 30 to 40 (sometimes),but the percentagesare rather low for the other activities. In order for the English classes to be conducted in English as a
prlnCiple, students need to use English as much as their teachers, and approprlate
measures should be taken.
I.2. The actual situation at UNIV
Sans specific statistical data on the actualsituation at UMV regarding theuse of English in college classrooms, the discussion will be confined topersonalexperiences and infわrmation gathered throughconversationswith people. To beginwith,it has been said that the classes most &equently taughtin English may be science classes at graduate school. There canbe at least two reasons for this. The &st reason is that among those graduate‑level science classes are fわreign students丘om other countries whose Japanese competency necessary for understanding lectures and discusslng
subjectmatters may not be atthe required level while theirEnglishcompetency is. The second reason is that it may not be so challenglng for Japanese teachers to teach their classes in English as the teachers and students use formulaeand English termir10logy specific to thesubject.At the undergraduate level, there are mainly two kinds ofEnglish classes: those for non‑English majors and for English ma)ors. Itcan be speculated that classes for English ma)ors are more &equently taught in Englishthan those for
Ron‑English ma)ors・Among the classes for English maJOrSarealso two kinds:
skill‑oriented classesand content classes・ Again, it canbe speculated that the former is
more &equently taughtin Englishthan the latter, but it stays as speculation. On the whole, however, the丘・equency of theuse ofEnglish at UW must be much lower than thatatJHS and SHS.Anin‑depth survey on theuse of English atUNIV isurgent.
2. Possible reasons forthe 一ack ofEnglishuse
lt has been shown that &om JHS through UNIV Englishclasses are not taughtin
English as much as they can be. What are the possible reasorlS for this? First,we look into the twelve possible reasons common to JHS through UNIVand two possible
reasons particular tomS and SHS.
2.1 JⅡS throughUNIV
Oyabu (2007)points out five possible reasons why the teachers ather college do not teach English classes in English, and four of those reasons can be the same &om JHS through UNIV.
(1)Internalization of "native speaker fallacy" and infTerioritycomplex: Some teachers tend to think that they should be at)leuse English justlike native speakers if they
were touse English in class.Asthey cannot, they do not use English inclass, just
like underachieving students who do not speak Englishas they are not good at English and have an infTeriorlty COmPlex.
(2)Absence of role models: Most Japanese teachers of Englishdid not have English
classes taught in E喝1ish by Japanese instructors when they were students; tberefbre, tbey have not bad role models.
(3)Traditionalstyle of instruction requiring a lot of talking on the part of the teacher:
Traditionally, teachers have talked a lot in class, and there may be teachers who think that they cannot give the same amount of information in English. Theyare not fTami1iarwith teaching methods where students talk much more thanthe teacher. In other words, they lack training ln teaching English throughEnglish.
(4)Lack or theoretical knowledge about SLA (secondlanguage acquisition):Due to a
lack of theoretical knowledge in SLA, some teachers may feel reluctant to teach English class in English.
Besides these reasons, Mizohata (2006),a teacher at a senior highschool,and Sato
(200S),a teacher atcollege, add the fわllowing.
(5)IneWICiency: it is inefficienttoteach in English as it is time‑consuming and students get00血sed.
(6)Preconceived idea about how English should be taught and learned by the students・
(7)Lack of experience for the teacher and students touse English among Japanese・
(8)Lack ofteachers'English proficiency
The fo)lowlng possible reasons can be added tothis list・
(9)Lack of teachers, confldenとc:teachers have competence butnot confldence.
(10) Lack of time: too much time is necessary in preparation and inconducting class itself if done in English.
(11) Beliefthat the instruction in Japanese is better thanthat in English: In the class taught inEnglish,students may not understand class content lOO% like in the class
taughtmainly in Japanese.
There may be other possible reasons, but it should be considered that these are the main ones.Also, as tothe absence ofroles models in(2),some may argue that there are native‑speaker teachers who can be role models, and they are certainly good role models of a kind,and Japanese teachers can learn a great deal &om them. Japanese teachers who teach EnglishinEnglishare another kind of role models as they share learnlng experiences of theEnglishlanguage with studentsand can be more realistic role models.
2.2 JSⅡ and S耳S
(1)Entrance?xaminations:If English classes are taught in English,there might be
apprehension that the necessaryknowledge and skills for entrance examinations
may not be obtained by students. This apprehension seems tobe greater in SHS than inJHS.
(2)Lack of uniformity: Inmost JSH and SHS, several teachers teach difrTerent classes of the same grade, and sometimes it is difficulttoarrive ata consensus to teach their classes in English.Also, in the same situation, ifsome of them teach their classes in English and the others in Japanese, there isa possibility that those teachers teaching in English cannot cover all the lessons planned to be finished by all the teachers including the ones teaching in Japanese, as it may take more time to teach thesame thing in English. This may leadsto reluctance to teach English in English.
3. Empirical evidence
ls there any empirical evidence that TETE ismore efficient than teaching English throughthe mother tongue? According to Eldridge (1996),"...there isno empirical evidence to support the notion that restricting mother tongue use would necessarily improve leaning efficiency. ‑." (p. 303).Also, I remember the following dialogue between myself and Prof. Tomohiko Shirahata atShizuoka Universlty, a researcher in SLA (personalcommunication).
Hayase.I Which ismore efficient, teaching English throughEnglishor teaching EnglishthroughJapanese?
Prof. Shirahata.・ I don'tknow. We need to experiment.
What Prof. Shirahata meant is that hewas not aware ofempiricalresearch atthat time to
answer yes or no. Final一y, Kanatanj (2004)writes that he has not heard of evidence
showing that themore Englishteachers use, the more students'Englishimprove (p.9).
If there isno depeJldable empirical evidence about the e用cacy of TETE, we need todepend on theories and hypotheses of SLA.
4. Theories and hypotheses and (heir叩PJication jn the chssroom
Very few people should disagree with the proposition that input plays the central role in SLA. The greatest part of language acqulSlt10n Occurs through the understanding of messages (Shirai,2008, p. 115).Sato (2009),aRer reviewingthirteen papers written by researchers, writes, "There isno lack of theories or hypotheses which regards input
as a precondition for leaming" (p.371).Finally, "as long (sic)there is input, acquisition will occur"咋ike& Young‑Scholten, 2009, p. 1).
How does Input WOrk in the classroom? First ofall, English used by the teacher is
a majorsource oflanguage input for his/her students (Mee‑Ling ,1996, p. 186)・Hughes (1978, p. 6) a)so contends that "…the classroom situation is a genuine social environment which allows 'the meaningful situational use of the language." In other words, teachers need to make efforts to maximize their students'exposure to English.
Secondly,anEnglish・environment, ln Which the use of Englishis abundant, helps the students to be convinced that it isan English class and that the conviction surely motivates them to study the language. Thirdly, through TETE, teachers become more
confldent in TETE,and atthe same time, they realize the unsatisfactory areas as to their English,and they work more to improvethemselves. In other words, unlessthey express themselves in English, they can't find what they can't express in English・ If they know what they can't express in English, then they can improve themselves by 丘nding approprlate StnlCtureS and expressions in books, dictionaries, etc・・For example,
on one occasion, Iwas trying tO explain what "meal" of "oatmeal" means・ The only word which came up ln my mind at that time was upushedn・ Iam sure the listeners understood what Imeant as I made gestures showing oatmeal as realia. But Iwas not satis6ed with my word, so later I consulted a monolingual dictionary and found out that the better word is "crushed." When I explain about the word "meal" in the future I will surely use this word learnlng &om my experience of using English.Students can also improve themselves in the same way. Last but not least, teachers cansave themselves from a double standard in which they encourage their students to use Englishadding
that it is all right to commit mistakes while they are conducting their classes in Japanese. Teachersalso do make mistakes for various reasons in spite Ofthemselves, but it shows to the stu.dents that it ISall rightto makemistakes in uslng the target language. One caveat is what Sato (2009)aptly says as follows: "we maintain the attitude thatwe are trying tobecome more accurate and nuent" (p.372).
5. SpecirIC Suggestions
Now that the importance and implications of TETE isfuIly awared, several
specific suggestions to promote TETE,with the understanding that the primary
objectiveofTETE isto improve the students'Ianguage, and topresent them with :
(1)carefully controlled listening practice and
(2)opportunities f♭r productive practice.
(Adaptedfrom Lewis & Hill(2002),p.12)
5.1 Decideto do it in Englishfrom day one
When teachers teach English throughEnglish,the Brst thing they should do isto decide to teach English throughEnglish.This is crucial. It is important to make a firm resolution to do it. In order totell themselves to put into this resolution into practice,1t is alsovital for them to be convinced that it is necessary to teach English through English.Allof them should have their own reasons why they teach English through
English. For me, not uslng English in class means depriving the students of opportunities to expose themselves toEnglishand using ltaS a means Ofcommunicatiop・
For many students, classrooms are the only situations where they can improve their skills in English. Teachers should not rob the students of the opportumty to improve themselves. They need to be firmly conscious of the importance of TETE by having theirown reasons f♭r it.
5.2 Break the psychologicalbarrier
One ma)or problem for many English teachers isnot their English skills. They have enoughskillsto conduct their classes in English, but theyare afraid of something.
They have hesitation・ They have anxiety. They have worries. They worry whether things go smoothly. They are reluctant to use English. They are afraid of making mistakes,
even though they say to their studeTltStO use Englishwithout being a丘aid ofcommlttlnB
errors as discussed atthe end of section 4. In other words, they have communicative
competence, but they may not have enoughcommunicative confidence. So, what they need more isnot English skills, but more encouragement to boost their confidence to
use English.See the listofencouragJng expressions about using English in appendix A.
5.3 t7se code‑switchillg
A dictionary definition of code‑switching goes as follows.・"a change by a speaker
(or wdter)斤om one language or language variety to another" (知chards,I. C. eta], ed.,2002, p.81).TETE does not advocate thatEnglishshould be usedall the time &om the beginnlng tO the end of each class, It allows judicioususe of the mother tongue.
There are occasions where use of Japanese ismore efficient than that of English such as grammarand vocabulary explanations or instructions for the procedure of task activities,
even thoughthereare scholars who argue that English should also be used in these occasions as they present genuine opportunities for teacher‑student communication in the classroom ¢Iarbord,1992; Moody, 2003).As mentioned earlier,也e MEXT's new
guidelines to be implemented in 2013also states clearly thatall the English classes at SHS should be taughtin English as a principle (Koko no gakushushidoyoryo kaiteian,
2009).We should, however, switch back toEnglish as soon as the necessary use of the Japanese language isover. See the checklist of code‑switching ln appendix B.
5.4 Improve ourselves
All teachers need to be challenged toconstantly improve themselves (Dash,2002, p. 15).Fukushima (2006,p. 193)agreeswith Dash and says that it isa basic need for English teachers to make continuous effortsto improve theirown Englishskills. It is also necessary to improve their teachings skillsas well・ The so‑called comprehensible Input represented in ui +1v is certainly necessary for teachers tohone their English skills, especially to develop nuency and automatlClty, but Hi +1乃surely does not contain too many unknown lingulStic items1‑grammar, vocabulary, or pronunciation for that matter‑‑as the formula showsI For English language acquisition for acqulrer teachers to
occur, however, itcan be argued that they also need ‖i +1+cL''which is English fTar
beyond Englishof their levels. It is difficultto define how much beyondthis "i +1+α"
should be, but one example of this for SHS students may be English studied for college entrance examinations・ This example may demonstrate the importance of ui +1+α''and it may also glVe SOme hint for teachers as to what kind ofEnglish they shouldaim at・Or many of them candefine theirown ̀̀i+1+αM based on theircurrent skills and their goals in English̲ "i +1+a" should make a world ofdiffTerence・
Teachers also need to improve theirEnglishin such a way that they can modify their English to help their students to understand them. EffTective skills in modifying
English are paraphrasing and making appropnate example sentences and phrases on the spot. Use of bodily movements like gestures and facial expressions, use of drawings, illustrations, and diagrams, and use of realia, pictures,and videos are natural tools to assist modifyingEnglish. These auxiliaries and the use of modified English constitute animportant part of teaching skills.
Conclusion
Teachers teach English f♭r the sake of their students. They want their students to improve theirEnglish.TETE should play a ma)or role in achieving this goal. Now that various important aspects of TETE have been understood, teachers need to work on
more specific methods and techniques to be used at JHS, SHS, and UMV while increaslng TETE and their English and teaching skills. Primary schools need to be includedinthe discussion as English activities with a long history at pnmary schools will be compulsory beginnlng in 201 1. Though the main aim ofEnglishactivities isnot teach English skills, English is used by the teacher and students. Itwill be necessary to discuss TETE at pnmary, secondary and tertlary Stages,and a close cooperation throughout the three stages will be crucial in order to make TETE bear expected results.
Referen(:es
Dash, P. S.(2002).Englishonly @0)inthe classroom: Time forareality check?
AsiaT7 EFL Jouma], December 2002. Ret^eved Jamury 3, 2009,丘om
http:/∧InW aSianW
Eldridge, J.(1996).Code‑switchingina Ttu.kish secondary school. ELTJouma), 50(4),303‑3 1 1.
Fuknshima, M. (2006). Research notes on teaching Englishin English. Joumal oflheFacu]o,
KokzLSal'Kyoyo
ofToyama LTniversjO,ofln(ema(iona[ Studies, 2, 187‑194. Retrieved January 3, 2009,丘om httpJnibmnWf
Hazbord, I.(1992).The use of mother tongueinthe classroom ELTJoumaL 46(4), 350‑355.
Hughes,G. S.(1978).A Handbook ofc/assTt)OmEng/ish. Oxford: Oxford University
Press.
Kanatani,K. (2004). 「オールイングリッシュ絶対主義」を検証するrIToveri&
"The absolutism of allEnglish''】.The Et7gh‑sh 7Teachers Magazine▲ 52 (1)),8110.
Koko no gakushushidoyoryo kaiteian高校の学習指導要領改訂案【Aproposal revision of thecollrSe Of study for senior highschool】.December 23,2008)Asahi Shi17bun, p. 2 1.
Lewis, M・ & Hill, J.(2002).Practl'ca/techniquesfoT・]angzLage (Caching. Boston: Thompson.
Mee‑ling, L・(1996).Arealityshock‑teaching Englishthrot1ghEnglishor Chinese? EdzLCafion JozLmaL ofHang Kong lnsLitzLJeofEducational Reseanh, the Cht'nese LTmlversjO,OfHong Kong, 24(2),
173‑191. Retrieved Jantlary 3, 2009,允om httl)J/sllrLZil.)ib.1血.hkntkio/view/33/33(伽2 1.p4f MEXT (2005).英語教育改善実施状況調査結果概要【Asununary ofthe血dings of the s町Vey Onthe
implementation of improvement of English education],faxed bythe MEXT, January 6, 2009.
MEXT (2006). 「英語教育改善実施状況調査」について【Aboul̀Ⅶle Survey Onthe implementadon of improvement ofEnglisheducation"],Retrieved January 3, 2009, bm
叫tT):/^vww.111eXt.gO.jDh mellu/toukei/001/06O322 1 I/OOl.htm
MEXT (2007).平成18年度小学校英語活動実施状況調査及び英語教育改善実施状況調査(中学 校・高等学校)について【Aboutthe suⅣey onthe implementation ofEnghsh activitiesatprimary
school and the implementation of improvement ofEnglisheducadoninthe academic year 18 of Heisci (juniorhighschooland senior highschool)】,Retrieved January 3, 2009,from
̲httl):/^nnv.me九t.RO.ion) menunlOudouノ1 9/03/07O3O8 1 i.lltm
NEXT (2008).平成19年度小学校英語活動実施状況調査及び英語教育改善実施状況調査(中学 校・高等学校)について【Aboutthe suvey on the implemenbtion ofEnghsh activities at p血叩 school andthe implementation of improvement ofEnghsh education inthe academic year 19 of
Heisei Omior highschoolandsemior highschool)] ,Retrieved Jantq 3, 2009,斤om
l刑p:〟ml叫/03/鵬03 I 920.htln
Mizohata, Y.(2006).Message toamailing list(旭ekeneigojugyokenkyukai:
mie eiivukellf(,halruliinla.CO.jt))[mie̲eijyuken:03 842]Angtut 23, 2006.
Moody, M. (2003).GraJnnuTinstruCtion in English in Japanese highschool.触eedi'7gS 0/(he 2gA AnmJa] Cbnvent71on
ofLheNatioTla[ Bhg[jsh E血cafio11 SbcieO,,pp. 649‑650.
Oyabu,K. (2007). 「英語で教える英語の授業」 :非母国語話者教員をとりまく状況分析と理論的
枠組み田nghsh chsses taughtinEnglish:A preliminaq study on norLnadve English‑sp血血g teachers'decisioninusing Enghsh asthe hngtnge ofinstrucdon].FoTm'gn LangyLqe EdzLCatjon Forum
ofthe FoTeigTI LangzLage IWtute
ofK‑azawa t[niver・siO,,), 26‑36. Retrieved January 3, 2009,血)m hltD二胤handle.ncu2297/I(鳩2O
Pike, T・ 皮 Young‑Scholten, M・ (2009)IIntroductilon・ In T. Pike & hL Yotng‑Scholten Phs.i InjWt
matters inSLA.佃p.1・26).Bristol: Multi1ingualhhtters.
Richards, I. C.etal..申d.).(2002).Longman D7'c(ionary ofLa"gyage TeachL'r7g & App)L'ed Linguish'cs, (3rd d.).London: Pearson Education.
Sato, a (2008).Let's do itin English: Suggestions formore successhl Englishclasses in Englishwith code‑switching. (Oralprcsentationatthe 38血alu岨I convemion of the Chubu Endish La皿gtuge EducatiorL S∝iety, June 29,2008.)
Sato, R.(2009).Sugestions for creating asucces血I input‑dch En由ishcJasswith alitde code‑switchirLg.
Joumot
oftheChubu Eng/ish Language Educatio〝 SbcieO/,38, 37 1‑378.
Shirai, Y (2008).外国欝学ぎの#学√虜二貢菅野得tぼ何か【Science of foreign language leaming:
What isthetheory of second langtnge leaming acquisition?I.Tokyo: Iwanamishoten.
AppeTIdix A:List o†encouraglng eェPreSSioⅠlS abollt using English
Theseare expressions I have collected over many years.
‑1Don't be a舟aid of making mistakes; be a丘aid of losing chances to improve yourself when you don't even trytouse English
1‑Give yourself permission tomakemistakes and learn &om them‑a tlPOn Studying a foreign language, &om aWeb site oftheUniverslty OfTexas
‑‑Dare to embarrass yourself.
‑Let us failcreatively.
‑‑Run lingulStlC risks in conversation!
I‑Profit&om mistakes!
‑‑"Second language learnlng adually necessitates the making or mistakes."
‑‑"Learmng isfundamentally a process that involves the making of mistakes."
The above two quotationsare丘om: H. Douglas Brown (2007). PrinctPle of
Language Leaning and Teaching (5thed.),p. 159 and p. 257.
‑‑"Ihaveknown failures. Those I haverun &om taughtme nothing." Jane Fonda (2005).
Jane FoJ血I坤Llfe So Far. New York: Random House, p. 573.
I‑"Ihave won 3,000 races, but I have also learned a lot &om the horses that did notwin.