This is to certify that the

thesis entitled

Learning Japanese As A Third Language

presented by
Shizuka Sakurai

has been accepted towards fulfillment
of the requirements for

A MA Linguistics

degree in

 

 

WMJW

Major professor

Date—,1, "[6 - y/K

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

LIBRARY 7
Mlchlgan State !
University 1

.fi

 

PLACE ll RETURN aoxmmmmbmum yourncord.
‘I’O AVOID FINES mum on or baton duo duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L usunaAnAMmmAwavEm-Iomnuylmon
W1

 

LEARNING JAPANESE AS A THIRD LANGUAGE

By
Shizuka Sakurai

A T111518

Submitted to
Michigan State University
in partial fulfillment of the requirement
for the degree of

MASTEROF ARTS

Department of Linguistics and Germanic,
Slavic, Asian and African Languages

1994

ABSTRACT

LEARNING JAPANIBE AS A THIRD LANGUAGE

BY

Shizuka Sakurai

The present study investigates bilinguals learning L3. The Korean and
Chinese subjects, who are bilinguals in their native language and English and
who are learning Japanese as L3, are compared with Americans studying
Japanese as L2 at Michigan State University. There are nineteen Asians and
twenty Americans as subjects who are chosen from Japanese first and second-
year levels. Negative questions and some phrases in Japanese which contrast
in word order in English are selected as the items for the tests. The main
questions were the eadstence of L2 transfer into L3 and a question: does
bilingualism have any impact on an individual's L3 learning? . The results
between two levels and two contexts, conversation and reading, are compared.
It was found that the recognizable number of errors reflect the transfer from
L2, English, into L3, Japanese. The paper points out one of the possible areas

where multilingualism works disadvantageously.

LJnimdusiinn.

In second language acquisition (SLA), there have been numerous
studies done and much discussion about language transfer (Gass and
Selinker 1983, for example). The main focus of such research has been on
the interference between one's first language (L1) and a second language
(L2), in this case the target language. The present thesis investigates a
phenomenon in learning multiple foreign languages, which has been an
understudied topic in SLA. It examines the L2 influence on the third
language (L3). Specifically, I attempt to answer the following question: " Is
there any language interference between L2 and L3 in L3 learning?" , and
to determine whether L2 is transferred into L3 Just as L1 transfers into L2.
Additionally, it addresses the question of whether bilingualism has any
impact on an individual's L3 acquisition. The answers to these questions
will give us a clearer picture of what is happening in a learner's head in
trilingual learning. I

As a result of the increasing internationalization and heterogeneity
of the population in modern societies, more and more people are becoming
multilingual. Japanese is one of the languages for which the number of
learners has risen rapidly worldwide in resent days. In the Japanese
classes at Michigan State University, too, there are many nationalities of
students represented. I chose American, Chinese, and Korean students as
the subjects of this study. I became interested in this topic as a result of
teaching Japanese as a foreign language. in the Japanese classes which I
teach, there are some Chinese and Korean students who have learned
English as L2 and who are studying Japanese as L3.

English is the medium of instruction in their Japanese classes as well

as in their learning materials, and I have observed that these students

2

make some of the same English-influenced errors in Japanese as do their
native English speaking peers. This was surprising because many features
of Korean grammar and some of Chinese grammar are similar to Japanese
and one would, therefore, expect the positive transfer from L1' to L3.
Actually we do often times find this kind of positive transfer from L1 in
their interlanguage. However some of their errors in Japanese reveal the
influence of English sentence structure. It appears that these errors arise
from their exposure to English on an everyday basis. Therefore, regardless
of the syntactic similarities and the expectation of positive language
transfer from L1 to L3, the transfer from L2 to L3 also appears to exist.

In the present study several grammatical structures and negative
questions were selected as the items to be asked. Chinese, Korean, and
Japanese all share similar properties in their structures, while English
does not. My teaching experience has indicated that the items are all
highly predictable sources of error for native speakers of English. The
study tries to confirm whether the same errors by Americans are
committed by native speakers of Chinese and Koreans. The data were
collected through oral and reading materials and grammatical judgment
tests. Background information on the subjects was also gathered. There
were four Chinese and fifteen Korean, and twenty American students, in
total thirty nine, who were enrolled in first-year and second-year
Japanese courses in spring semester in 1994..

Contrary to the common notion that it is easier for a bilingual to
acquire a third language than it is for a monolingual (Cenoz 1994), my
hypothesis is that knowing another language sometimes can be a
disadvantage in learning a third language. When one has to use some

structure in one language, a few different structures in different

3

languages may merge in one's head and one may choose an inappropriate
structure. Why such an inappropriate structure from one language is
chosen is also an interesting question to examine. In this study, I view and
approach "language transfer" from a wider angle, that is, from that of

knowing several languages.

LErexiQuumdies.

Discussion on transfer started from Iado's claim in 1957 [p. 25]:

We know from the observation of many cases

that the grammatical structure of the native

language tends to be transferred to the foreign

language. The students tend to transfer the

sentence forms, modification devices, the number,

gender, and case patterns of his native language.

We know that this transfer occurs very subtly so

that the learner is not even aware of it unless it is

called to his attention in specific instances.
His claim had a great impact on the field of second language acquisition.
Since then, various studies have been conducted to prove the existence of
transfer, and to answer questions such as: what can be or actually is
transferred?, how does language transfer occur? What types of language
transfer occur? For instance, Schumann (1979) researched the acquisition
of English negation by native speakers of Spanish and compared it to
speakers of other languages like Japanese, French, German, Norwegian,
Taiwanese, Greek, and Italian learning the same form. His finding was that
the position of negation tends to be determined by the position in L1,
either pre-verbal or post-verbal.

However, this notion of transfer has been challenged by the
important research of linking L1 and L2 acquisition. "L2 - Ll hypothesis"
was questioned first by Selinker (1966). Then, a paper by Dulay and Burt

(1973) exemplified this. Dulay & Burt argued that Chinese and Spanish

4

children acquired English morphemes the same way and said [p. 256]
"there does seem to be a common order of acquisition for certain structures
in L2 acquisition" . Therefore, they denied the difference which comes
from different L1 and stated that L2 is acquired as L1 is acquired.
Furthermore, Bailey et a1 (1974) did research using seventy—three Spanish
and non-Spanish speaking adult subjects learning English as a second
language and reached the same conclusion as Dulay and Burt did. That is,
there was no L1 transfer in L2, supporting the "L2 - L1 hypothesis". They
claimed that errors are not transferred but are developmental, caused by,
for instance, overgeneralization of rules. 7

However, the issue is not so black and white. We cannot decisively
declare what is the source of errors. It may be due to transfer from L1 or
other languages, or it may be due to developmental errors. According to
George (1972), one third of errors were attributed to L1 transfer, while
Dulay and Burt ( 1975 cited in Gass 1979 ) found less than five percent of
errors caused by Li transfer. They all implied the existence of the source
of errors other than L1 interference. Major (1987) investigated the
interrelationship of interference and developmental elements in L2
phonological acquisition, and while admitting considerable evidence for L1
transfer in L2 learning, he said [P.120]:

Interference processes first predominate and then

decrease over time. However, developmental processes

are at first infrequent, but later they increase and then

decrease in frequency.
His theory of the relationship of interference and developmental processes
to time is described as the "Ontogeny Model" and is illustrated by the
following figure [p. 103]:

frequency . fl

 

 

 

 

interference developmental

He claims that this model for phonology may also apply to syntax. Taylor
(1975) presented a similar model in syntax: beginning learners have more
interference errors than intermediate learners for whom developmental
errors such as overgeneralization are more common. It may be that since a
learner does not know much about the target language in the beginning
he/she relies on his/her L1 more than he/she does later. However,
identifying the source of errors is actually not an easy task and how to
account for the criteria for transfer and developmental errors is
problematic. As Zobl (1982) stated, transfer and developmental influences
interact with each other and it is hard to differentiate them. We cannot
ignore the ambiguity with regard to this problem (Gass 1984).

Despite the above views to the contrary, it would be unwise for us to
deny the phenomenon of language transfer. In addition to other possible
sources of error, my teaching experience has led me to believe that
language transfer does exist and I agree with Gass and Selinker (1983) in
that [p. 7]:

We feel, however, that there is overwhelming

evidence that language transfer is indeed a real

and central phenomenon that must be considered

in any full account of the second language acquisition
process. '

6
language transfer needs to be investigated as an important phenomenon

in order to better understand language acquisition in general, human
capacity for language, and for its pedagogical implications. later in this
thesis, I shall return to this issue and relate it to my research.

Below, I discuss some other notions which are related to language
transfer, which may give some insight about the present project. Transfer
is said to be determined by the distance, that is, similarities and
dissimilarities between L1 and L2, from which the concept of contrastive
analysis came out. In the contrastive analysis, it was believed that errors
can be predicted and reduced by a systematic comparison of any two
languages. The structures of L2 that coincide with corresponding
structures of L1 are assimilated with great ease as a result of positive
transfer. Having an L1 which is closely related to L2 can be advantageous
in many ways, including the carryover of L1 items and structures to the L2.
On the other hand, contrasting structures present considerable difficulty
and give rise to errors as a result of negative transfer or interference
between two languages that are distant. Also, a lack of congruence
between the L1 and the L2 may result in slower development of the target
language, or it may cause fossilized productions. Keller-Cohen (1979)
conducted eight-month longitudinal study about Japanese, Finnish, and
Swiss German children acquiring the question formation in English. She
found that the different developmental patterns are able to be explained by
the similarities and dissimilarities between English and their native
languages. The slow development of the productive use of yes/no questions
in English on the part of the Finnish child in comparison with German and
Japanese children was found. Also, an interesting observation concerning

transfer issue has been made by Ulijin and Kemmpem (1976, cited in

7
Krashen 1983), that is, L1-L2 differences do not appear to present a

problem in reading. If this is true, then errors caused by Li transfer
should not be found in understanding of written text which requires more
attention, as often as in conversation which requires quicker response.

As lado (1957) states, transfer occurs unconsciously and students are
not aware of doing so. Krashen (1983), however, states that errors by Li
transfer can be reduced or "cured" by monitor use, or in carefully
monitored speech. He relates his theory to the "cure" of language
transfer. His monitor theory has to do with an effort in conscious language
learning made by the learner to control his language production and to
self-correct it. Krashen's hypothesis is [p.136]:

Monitor can repair errors caused by

the use of the first language in cases where the

rule is learned but not yet acquired. This cure, however,

is only short-term.
The monitor use not only repairs interlingual errors, but also plays a role
in avoidance of some structure rather than troubling to repair it with the
monitor when a rule has been learned but has not yet been acquired.
Schachter (1974) first mentioned the phenomenon of avoidance clue to
difficulties, which can be predicted when the two languages are compared.
She also showed an example of avoidance of relative clauses in English by
Chinese and Japanese learners of English. However, as Krashen points out,
there are individual differences among learners and they may or may not
be effective monitor users. If the monitor is "on", it facilitates the
accuracy of items which are not yet acquired.

While what has been usually discussed in the literature is the

transfer from L1 to L2, Mackey (1968) addresses unique interference cases

from L2 to L1. He reports several phenomena in which various features in

8

the foreign language actually replace features in the native language.
This suggests that transfer does not necessarily have to be in the direction
from L1 to L2. It can‘ be bi-directional.

A second language refers to (larsen-Freeman and Long 1991 p. 7)
"any languages other than the language learner's native language". '
Therefore, with this broader definition in mind, we can expand the notion
of transfer from the conventional one (L1 to L2) to that which includes
transfer from L1 to L2, L2 to L3, and L1 to L3.

When examining language transfer in multilingual cases, it soon
becomes obvious that not many previous studies have been done. language
transfer by multilingual speakers seems to be regarded rarer than by
monolingual speakers. However, there are many countries in which
multilingualism is quite common. The study of language transfer in the
multilingual context seems to be an important but as yet understudied field.
There is a need for more research in this area.

Research in multilingual context has been done by some scholars,
such as Magiste (1979, 1986) and Baike-Aurell and Lindblad (1982 cited in
Magiste 1986). They all tried to answer the question: "Is learning a third
language easier for bilingual immigrant students than learning a second
language is for monolingual students?" In Balke-Aurell and Lindblad's
study in Sweden, Grade eight immigrant students who are bilingual in
Finnish and Swedish are compared in a standardized test in English (L3)
with Swedish monolingual students. Among the bilinguals, there are two
groups: group 1 are those who use L 2 (Swedish) dominantly at home and
group 2 are those who use L1 (Finnish) dominantly at home. The result
was that group 1, passive user of L1, performed clearly better than group 2,

active user of LI, and group 1 did slightly better than group 3,

9
monolingual Swedish students. In Magiste's study (1979), a similar result

was obtained. In his research, the L1 was Swedish, L2 was German, and L3
was English and a group of seventy—four German/Swedish bilingual high
school students are compared to a group of trilingual students and
German/Swedish monolinguals. He found that [p.116] "trilingual subjects
needed more time to perform the more tasks in their second and third
language than bilingual subjects did in their first and second language".
He argued that monolingualism and passive bilingualism seemed to be
facilitated in learning a third language, while active bilingualism may
delay L3 acquisition. The finding goes against the common idea that the
third language learning is easier in general. The longer reaction time of
multilinguals may be explained as an effect of interference of competing
language systems. Kovac (1965) stated that [p.158-159] "real
multilingualism may be a disadvantage in all situations in which extremely
prompt responses are required to verbal stimuli,..." Therefore, it seems that
for an active two language user, learning a third language is not as easy as
was thought to be the case.

In the studies by Magiste, Balke-Aerell, and lindblad, they examine
whether active use of L1 influences L3 acquisition in a negative way or
not. Here, being an active bilingual is found to be disadvantageous.
However, in Jaspaert and Lemmens' study (1989), in which grade school
children in Brussels whose Ll was Italian and L2 was French were learning
Dutch as the target language (L3), they found that bilingual children do
not show much difference from monolingual children in the results of L3
tests. Also, being bilingual did not impede acquisition of L3 or a successful
transition to secondary education. This is supported by a Wagner and

Spratt (1989)'s investigation in Morocco that the instruction in L2 did not

10
hinder L3 acquisition. In addition, Cenoz and Valencia (1994) tried to test
the effect of bilingualism (Spanish and Basque) on L3 (English) acquisition
and also the effect of instruction in L2 on L3 learning that was a minority
language. Their findings were more positive than the findings in Jaspaert
and Lemmens' study (1989). One of the results found by Cenoz and Valencia
is that both being bilingual and using L2 as a medium of instruction
facilitate positive outcomes in L3 acquisition because [p.207]:

...they (bilingual students) use their knowledge of two

other linguistic systems (Basque, Spanish) and compare

them to the new code. This metalinguistic awareness could

possibly account for a higher level of linguistic competence

in a third language.

In addition, he attributed these positive outcomes to the subjects' "very
strong, positive attitudes toward their language." [pp. 206-207] . In
actuality, a success of second and third language learning is bound to be
affected by various conditions like the languages, ages, motivation, the
environment in which the language is used, and so forth. We have to view
various cases of acquisition from several angles in order to reach a better
understanding of second and third language acquisition.

Finally, Akiyama's (1979) study concerns the acquisition of negative
questions by American and Japanese bilingual and monolingual children
in the United States. Their mean ages were four to ten for Japanese-
English bilinguals, five for English monolinguals, and four to eleven for
Japanese monolinguals. (I shall explain the system of Japanese negative
questions in more detail in the following section, W
W ) Negative questions are answered differently in
English and Japanese. The difference is shown in the table below cited in

Akiyama [p. 488].

11

 

 

Table 1
Negative question Answer Underlying
speaker . intention
English Aren't you going? Yes, I am going.
speaker Aren't you going? No, I am not going.
Japanese Aren't you going? No, I am going.
speaker Aren't you going? Yes, I am not going.

 

In English one answers "yes" or "no" depending on one's intention about
the matter being asked, while in Japanese one affirms or denies
questioner's assumption by "yes" or "no". Therefore, the assumption is that
the person being asked might not go, and in Japanese "yes" means that
"what you seem to be assuming is right, I'm not going" whereas "no" means
that "what you seem to be assuming is wrong, I'm going". Akiyama's
question was which of the two linguistic systems was acquired earlier and
whether the two systems work independently or interactively in Japanese-
Engllsh bilingual children. He found that the two systems in answering
negative questions interact with each other in bilingual children, and that
the English way of answering was acquired earlier than the Japanese way.
It is interesting that, when the bilingual children were asked negative
questions in Japanese, a high percentage of them used the English
answering system. What Akiyama found has a great implication for the
present study in terms of English (L2) transfer to Japanese (L3) in
answering negative questions, although my subjects are adults and they

acquired English as their L2.

W
In this study, I selected syntactic Japanese structures that are

negative questions and four phrases, which Chinese, Korean, and Japanese

12

share, but which English does not. Hence, they are all highly probable
sources of errors for native speakers of English. As for the equivalents in
Chinese and Korean, it was confirmed in some grammar books, references
and by my informants 1 that basically their arrangement of words was the
same as that in Japanese. The Chinese, Korean, and Japanese structures
were formed one group, while the English structure was formed as the
other. If Chinese and Korean subjects make errors using English word
order, then we may assume that there is English influence. 2

The items treated in this study are classified into several types:
negative questions, time related expressions and modifier-head phrases.
First, as for negative questions, Table 1 from Akiyama (1979) in the
previous section shows the basic contrast between Japanese and English
systems of answering. Additionally, according to Martin (1962), Japanese
"yes" and "no" in negative questions can be analyzed as follows:

The words hai (oree) and lie are used to mean 'what

you've said is correct' and what you've said is incorrect'.

So if you state a question in a negative way, the standard

Japanese answer turns out to be the opposite of standard

English 'yes' and 'no', which affirm or deny the FACT

rather than the STATEMENT of the facts. lpp. 364-365]
llowever, Kuno (1973) views Martin's explanation to be too simplistic and
he classifies negative questions roughly into two patterns: one is a "neutral
(negative) question" in which a questioner is expecting a negative answer

from the listener or a questioner has no presupposition about what the

listener is going to answer [p. 274]. The other is "one that contains the

 

1 The informants are native speakers of each language whose major is
linguistics.

2 Although this assumption is possible, there still remained a question: it is
not very certain that Korean and Chinese students have actually mastered
the system of negative questions in English. Sometimes even native
speakers of English may make a mistakes. It may be interesting to
investigate how native speakers of English would respond to this kind of
formal tests of negative questions in English.

13

questioner's expectation of a positive answer", which is like "an
affirmative sentence with a tag question at the end". If one is asked the
former "neutral (negative) question", the answer is the one that Martin
defines. HaLUYes") corresponds to the English "no", while liefNo")
corresponds to the English "yes". However, regarding the latter negative
question with a positive expectation, since the questioner's assumption is a
positive one, as in English, haL is used for an affirmative answer and tie
for a negative answer. The following are some examples:
(1) a. "neutral (negative) question"
A. kinoo toshokan e ikimasen deshita ka.
yesterday library to go-not did
"Didn't you go to the library yesterday?"
B. Ha i, ikimasen deshita.
yes go-not did
"N o, I didn't' go."
Ii e, ikimashita.
no go—did
"Y e s, I did."
(2) b. negative question with a positive expectation
A. Gakusee ja arimasen ka.
student be-not
" Aren't you a student?"
B. Ha i, soo desu.
yes right be
"Y e s, i am"
11 e, chigaimasu.
no wrong
"N 0, I'm not."
Kuno says that the cues to distinguish the two types lies in the intonations
or nonlinguistic environments such as conversational contexts. In the
present study, negative questions with a positive expectation as in (2) was

regarded as a tag question and was eliminated. To make a contrast in

14

systems between Japanese and English clear, I used only neutral (negative)
questions as in (1).

In Chinese and Korean, the system of answering negative questions
is basically the same as in Japanese. According to Introduction to Spoken
Korean by Home and Yun (1950, p. 23), "..., when you use [1: (or ye) 'yes'
and £11119). 'no' in replying to NEGATIVE question, they are used just the
opposite from English 'yes' and 'no'". In Chinese, Mandarin Primer An
Intensive Course in Spoken Chinese by Chao (1957) says that "If the
question is in the negative, then the answer in Chinese will seem to be the
opposite to that of the English". As these materials are old, I asked some
native speakers of Chinese and Koreans (whose major is linguistics). They
all agreed with the statements in these books. Thus, it is possible to group
Chinese, Korean, and Japanese languages with regard to the ways of
answering negative questions.

As for the rest of the items treated, namely, time related expressions
and modifier-head phrases, the examples used in this study are given in the
following table.

 

 

 

Table 2
English Japanese
time related (1) 5:00 PM gogo go-ji
expression PM S-o'clock
(2) three times an hour ichi-ji-kan ni ichi-do
l-hour per 3-dme
(3) May 2, 1994 l994-nen go-gatsu hutsuka
l994-nen S-month 2nd
modifier-head (4) restaurant named Kyoto Kyoto toiu resutoran

expression Kyoto named restaurant

 

15
As seen in (1) gQgQ_gQfl_ ("5:00 PM"), examples were chosen in which the

word order in English and Japanese was contrastive. In Japanese, gogo
("PM") is placed before the time, M("5:00"), whereas they are in the
opposite order. In modifier-head expression as in (4) Warm
("restaurant named Kyoto"), in Japanese, a modifying phrase comes before
the head noun which is the opposite of the order in English. Therefore,
1510191919. ("Kyoto named") is placed before eega, ("movie") , and it
modifies eega, In this sense Japanese is a left-brancing and head-final
language. Basically, the word order in the Japanese examples in Table 2

also is the same in the Chinese and Korean equivalents.

$.Methsxi
We.
total

Korean students studying first-year Japanese ------------9
Chinese students studying first-year Japanese -------- 4 19
Korean students studying second-year Japanese --—-—-6
American students studying first-year Japanese ------— I3 20
American students studying second-year Japanese-------7

There are nineteen Asian students, consisting of nine Koreans and four
Chinese in first-year Japanese (in total, thirteen), and six Koreans in the
second-year course, which nearly corresponds with thirteen Americans in
the first year and seven Americans in the second year. There are twenty
female and ninteen male subjects. They are all students at Michigan State
University. All the American students are English monolinguals, whereas
Chinese and Korean students are bilinguals in their native languages and
English(the degree to which they are bilingual must vary). At the end of
experimental session the Asian subjects are asked to complete a

questionnaire about their English proficiency, language background,

length of residence in the U. 8., languages spoken outside of classes, and

16

their study of Japanese. The following table presents the summary.

 

 

 

 

 

 

 

Table 3
Korean Chinese Korean
(1 st-year) Q st-xear) (2nd-year)
average length of stay 2.4 2.1 4.7
in the U.S.(yrs.)
TOEFL score (average) 569.2 581.5 571.6
96 of time speaking
(outside the classes)3
EnglishL 28.8 57.7 49.5
Japanese 7.2 3.5 7.5
Native lg 63.3 38.7 45
96 of students who think 66.6 2 5 50
studying Jap.
in Eng. is inefficient
use of native language in 88.8 50 83.3
studying Jap.(96 of students)

 

 

 

 

 

The questionnaire used in this survey appears in the-appendix.
smegma

The data were gathered from five kinds of tests. Word order question
and the way of responding to negative questions are investigated each in
an oral context and in a written context. The first part of the four tests
concerns negative questions in oral context, and the second part in a
written text. In the third part, the word order of some expressions (see p.
14) uttered orally is studied, and the fourth part treats word order in a
written text. The oral context prompts the subjects to have a quick
response. On the other hand, the written text allows them to have enough

time to think. American and Asian student's responds were compared, as

 

3 We have to admit that this percentage is subjective estimation by the
subjects.

17

well as the difference between the first-year and second-year levels. The
fifth part is designed only for Koreans and Chinese students to attain
general information about their studying in the U. S. The outline of each
part is summarized in the next page.

Part 1: 8-to lO-minute individual conversation in Japanese
including 8 to 10 negative questions

Part 2: Choosing the right answers to 4 questions (including 2
negative questions) following a reading passage

Part 3: Explaining a schedule in Japanese, which is given in
English. The task includes the target phrases which
contrast in word order between Japanese and English.

Part 4: Grammatical judgment task. Students read sentences, which

include 2 grammatical and 2 ungrammatical sentences,
and answer whether they are grammatically right or
wrong.

Part 5: General information about studying in the U.S. (only
Asians)

In the first part, students had an eight to ten minute conversation
in Japanese with me. The topics of the conversation were generally about
their spring and summer breaks, sports and music, trips, and so on. In the
course of the conversation, I tried to ask eight to ten negative questions in
a natural context. The type of negative question is limited only to "neutral
(negative) question", as defined by Kuno(1973), in which a questioner is
expecting a negative answer from the listener or he/she has no
presupposition about the listener's answer. In this case, the responses
make a contrast with those of negative questions in English. The
conversation was recorded.

Secondly, I asked students to read a short passage which consisted of
four sentences. The passage was the one that even first-year students are

able to read and comprehend. Therefore, most of the subjects had no

difficulty in understanding it. The passage was followed by four questions,

18

including two negative questions. The other two questions served as fillers.
The students were asked to read them and choose the right answer from the
given choices. The translation of the passage is "I went to Japan with my
friend last year. I wanted to go to Mt. Fuji. But since it was snowing hard,
neither of us went there. That was too bad. " The translation of questions
and choices of answers are:
(1) Did this person go to Japan by himself?

1. Yes, he went there by himself

2. No, he went there with his friend. (the right answer)
(2) Did this person go to Mt. Fuji?

1. Yes, he did.

2. No, he didn't. (the right answer)
(3) Didn't this person want to go there?

1. Yes, he did.

2. Yes, he didn't.

3. No, he did. (the right answer)

4. No, he didn't.
(4) Didn't his friend go there, either?

1. Yes, he did.

2. Yes, he didn't. (the right answer) _

3. No, he did.

4. No, he didn't.
The choices include both English and Japanese ways of answering to test
which answer the students would choose. In addition, the question (3)
requires students to affirm the thing that the question concerns, while (4)
requires them to negates it.

In the third part, the subjects were asked to skim the following notes

written in English and to explain it to me in Japanese:

19

 

Please explain in Japanese what you did on Sunday
1:00P M got up
2:00 -4:00 stayed home because it was raining
6:00 went to a restaurant named Kyoto
9:00 called my parents three times an hour
2:00A M went to bed

schedule of March 15, 1994

 

The bold phrases are the tested phrases chosen for this study. They were
not in bold on the note given to the subjects had these notes. The four
items in bold form the opposite word order in Japanese and it is intended to
test whether any English influence on the word order in their Japanese
could be observed in a speaking context.

The fourth part also concerns testing of word order, but in a written
text in Japanese. The students read sixteen sentences, in which each
sentence contains one of the same four items of word order as the previous
part has. Each item has four sentences that consist of two grammatical and

ungrammatical sentences. For example, two sentences have the phrase

appearing as W ("5:00 AM/PM") and two other sentences
with WWW/PM 5:00") in Japanese. All the sentences are in

random order and they are judged on a four-point scale: wrong, probably
wrong, right, and probably right. This four-way distinction serves to
indicate their degree of certainty. In order to exclude extraneous factors, I
asked subjects to underline a place that they think is ungrammatical or
probably ungrammatical. If they marked phrases that do not concern any
of the four items that I am studying, I eliminated them from the data so that

the data would be based on the students' knowledge about the four items

20

only. The subjects were told that there were no mistakes in particles,
characters, and so on. They should not go back to the previous sentences to
change the choice that they made. The actual sheet that was used for this
test is in the appendix. .

The fifth part was filling out the information sheet. A brief
summary of the chart appeared in "3. 1. subject" section. The sheet that I
used is in the appendix. This part was added to obtain the students'
background information about studying Japanese and English and about
languages they use in daily life, which may give some insight when

interpreting the data.

ELEM
Weizmann.

The results of each parts are summarized in the following tables.
First, Table 4 shows the results of negative questions in conversation. The
table illustrates the background of the subjects: first-year and second-year
levels, and Chinese, Korean, and American. Five ways to answer negative

questions were found. The correct ways were, for example, either :9,
Whig ("Yes, I didn't") or WVNonid") They
sometimes answered incorrectly, mm ("Yes, I do") or jig
WC'NO, I don't") which are the same as English way of
answering. Additionally, some subjects did not say clarify "Yes" or "No"
overtly and just said W01 did") or W01 didn't").

Thus there were a total of five types of answers.

21

.3ch A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3:62 >. 5 :6. in Saopm >. H: mam. 5% 20 «\z HOS—a
<8. H 905.". 20. H do :2»: <8. H do 20. H don}. :08:
g . ~8me Nu Hm $30.33 w Hm HENHboS Nafimbes mu
:34.
0338 Hw a H03 Ewes A w .2 Hmbes H 23.33 mu
$34.
534. 88— AH Ho moGobé u Hm ”$30.33 33933 H~o
£9.85 an 05:83
x08»: Hm w 33.9.33 w Ho HoANmues Nmauuos mu
Nan v4.
Hm" an NEH UR mu 3 $13.38 Ho we fiwauos $23.33 Hmu
8an
. genome Swan»: 3 No «23.33 HN No 33933 H2 Hmhgv HNH
534.
>324an H H m Hofimuos m m H2 $.83 wwiqhos 3
Manda.
Hun an NEH 5 mm we. «2.3.88 No Nm Amfimhoe mHANuHoS Hg
88.—
88— H HN do 323.38 we 3 33933 H HNH ghee w;

 

 

22

Five types of answers:
1. Wrong answers (ways of answering in English)
(1) Yes, I do. .
(2) No, I don't.
2. Correct answers in Japanese way (ways of answering in Japanese)
(3) Yes, I don't.
(4) No, Ido.

3. Other
(S)Answers without overt Yes or No

Table 4 is the summary of the result in part 1.
In this test, similar results are found among Americans, Korean, and
Chinese. About half of the answers are grammatical, which follows the
Japanese pattern, while about one fourth of the answers are
ungrammatical, which uses the way of answering in English. Although we
have to admit the limit of this study that there are only two levels of
students, first and second and the number of subjects is not very large, a
few differences between two levels are found. At the higher level, fewer
correct answers and more incorrect answers with the English pattern are
obtained for Korean subjects. (However, this may be due to the fact that
first-year students had the instruction of negative questions more recently
than second-year students.) Additionally, at the second level, the
percentage of self-correction increases among Koreans and Chinese and
the percentage of answers without Yes/No increased among both
Americans, Koreans, and Chinese.

Table 5 presents the result of Americans that is partially taken from
Table 4. First, regarding American subjects, it turns out that the
percentage of the correct answers is not very high and that about one
fourth of their answers are marked wrong in the English way. The
category called "self correction" is for the answers that is corrected from

the English way to the Japanese way. For example, they first answered by

23
saying" No, I didn't", realized that they made a mistake, and changed it into

"Yes, I didn't". The numbers of self correction are included in correct

 

 

answer.
Table 5
correct A. in self wrong A. in no Y/ N total
Jap. way correction Eng. way

American 70 (57.896) 8 (11.496) 32 (26.496) 19 (15.796) 121
lst yr.
American 19( 28.396) 0 (096) 16 (23.896) 32 (47.796) 67
2nd yr.

total 89 (47.396) 8 (8.996) 48 (25.596) 52 (27.696) 188

 

x2 = 2.355 P-= 0.125 > 0.05

The self correction is more frequently used by the first-year American
subjects than by the second-year Americans. Instead, the second-year
American students avoided stating "Yes" or "No" overtly, which is explained
by a large increase from 15.7% by the first-year students to 47.7 96 the
second-year students in "no Yes/No category". With this increase, there is
a decrease in second-year level students in the percentage of correct
answer: from 57.8% (first-year level) to 28.396 (second—year level).
Therefore, it seems that this large decrease is explained by the large
increase of avoidance in "Yes/No" usage. Iiowever the Chi square test does
not show a significant difference.

Table 6 compares American and Asian, Chinese and Korean, students.

24

 

 

 

Table 6
correct A. in self wrong A. in no Y/N
total
Jap. way correction Eng. way

American 89 (49.096) 8 (8.996) 48 (25.396) 52 (27.596).
lst 82 2nd yr. 121
Asian83(44396) """ 7' 13.192?""'74';'('2'5.'s'<36"""8’613'27660 """
I238: 2nd yr.
x2 - 0.004 P - 0.998> 0.05

Similar results among Americans and Asians were obtained as shown in
Table 6. The Chi square test 4 also shows that they are not significantly
different. This result can support my hypothesis that there may be a
second language (English) interference in a third language (Japanese)
learning. Chinese and Korean students have a similar tendency to use the
English way of answering negative questions. American students show
the higher percentage in answering in the Japanese way than Asians,
while Asians make use of the avoidance of saying "Yes/No", which is
sometimes found also in Japanese.

Table 7 shows an unexpecting result that, at the higher level, fewer
correct answers and more incorrect answers were obtained. In addition,
the percentage of self correction increases. One fact that the percentage of
"no Yes/No" increases in the second-year level, from 29.1% to 37.396, may
due to the students purposely avoiding "Yes" or "No" and come to utilize
avoiding saying "Yes" or "No", although not saying Yes or No in Japanese is

not incorrect but rather typical.

 

4 The Chi square test shows the relations between "correct answers in
Japanese way" and "wrong answers in English way".

25

 

 

 

Table 7
correct A. in self wrong A. in no Y/N total
Jap. way correction Eng. way

Asian 60 (50.096) 4 (6.696) 25 (20.896) 35 (29.196) 120
lst yr .
Asian 23 (34.396) 3 (13.096) 19 (28.396) 25 (37.396) 67
2nd yr
x2 = 3.110 P== 0.078>0.05

The comparison between Chinese and Korean in the first-year level
is presented in Table 8. The figures for the percentage of each category
turn out to be similar. The Chi square test also shows that they are not
significantly different. Therefore it can be concluded that Chinese and

Korean students tend to perform in a similar way in negative questions in

 

 

 

conversation.
Table 8
correct A. in self wrong A. in no Y/N total
Jap. way correction Eng. way
Korean 41 (49.396) 3 (7.396) 18 (21.696) 24 (28.996) 83
1st yr.
Chinese 19 (51.396) 1(5.296) 7 (18.996) 11 (29.796) 37
lst yr .
x2 -0.112 P=0.738>0.05
5 2 I! ll !' l 1.

The next page shows an overall table of the result in negative
questions in the reading task. The subjects first read a passage in Japanese
and answered four questions including two negative questions. The

students had four multiple choices in each of two negative questions, in

26
which there are two choices of Japanese way and two of English way. Of

the two choices of Japanese way, only one is the correct answer to the
question. Therefore, in Table 9, all of the answers are categorized in two
ways: "Answers in Japanese way" and "Answers in English way". Then,
each category is divided into two again: "Understanding" and "not 1
understanding". " Understanding" means that the subjects understand the
reading passage and question, and they are right with regard to the
content, for instance, in the "he, did" part, but they may be wrong
grammatically on the "Yes/No" part. For example, in a negative question
(3):

(3) fuji-san e ikitakunakatta n desu ka

Mt. Fuji to go wanted-not that
'Did this person not want to go there?’

1. ee ikitakatta desu.
yes go wanted-not
'Yes, he did.‘
2. ee ikitakuarimasen deshita.
yes go want-not did
'Yes, he didn't.‘

3. lie ikitakatta desu
no go wanted
'No, he did.‘ (the right answer)

4. lie ikitakuarimasen deshita.
no go want-not did
'No, he didn't.’

27

.3200

 

 

 

 

 

 

 

 

 

 

 

 

 

 

>393 3 H3958» tum >533 H: mamHHa. ink 83..
5.53265" 39 5:338:95 E :9 Eng
3. 3 d E 3 a. 3 fl 3. H...
20. :0 n5. 4.8.36 90:... <37» 29: zot H6 2a. <3. :0 En. 20. H3 995.. 20. H! 95:, <8. so 95.
3...»: £033 u o Ha H o H Hm u c H O o c u Ha
Ha. H4. HH.H8 Put «who. Hob: o8 Heb:
0350» H H w H o H A N H m c c o w H
Hue we. 583.. IR: uHHl the. oi $.30
SE. a H H HH N o H Ho u H a o o o o nu
Her .44. am: at H08 Net 98 ~58
7.033.. m o m o o o u H c H o o o H o
55. K... «mu: 03 «PH... 5.0! at Hard:
88H H H H H - N o N E. o H H o o o H u H
Ha. an HRH. 3.3 9.3. HHhi 3:3, 9H. 3.3..
2.5105 >388»: H 3 HH H o H Hm u u a H o H H Nu
HA. H4. 23. ..8 H~t ~58 .3. Not
23135 H H Ho 0 o c 8 A o A o o o a He
HRH. 3.. 3.58 08 HHS: Ham: 3. Hub:
88. Ho HH NH H o H 3 H u Ho H o H H H we
Hm». an Nun. 938 NM: 3.3.. Hub: Nut Hub:
8.». - .3 mo u o w mw Hw .H HH H o H Ha HH
F H98! ski Hire: It i 1 3H! Va: “Hub:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

28
2. "Yes, he didn't" and 3 "No, he did" are labeled as "Answers in Japanese

way", while 1. "Yes, he did" and 4. "No, he didn't" are labeled as "Answers in
English way". Of the choice 2 and 3, 2 is correct in the form, "Answers in
Japanese way", but not in the content. Therefore, it is considered "not
understanding". 3 is right both in the form and content, which is .
categorized as " Answers in Japanese way" and "understanding". As for 1, it
is wrong in form, "Answers in English way", although it is right in the
content, "understanding". 4 is categorized as 'Answers in Japanese way"
and "not understanding". Therefore, there are four categories, and each
category has two boxes for question (3) and (4) in Table 9.

Some findings in part 2 in reading task are similar to those in part 1
in conversation, although part 2 obtained lower percentage of influence
from English shown in "Answers in English way". in part 2, American and
Asian subjects present similar percentages in "Answers in Japanese way"
and "Answers in English way", that is, Asian students have the similar
tendency to use the English system of answering negative questions in
Japanese. In the second-year level, American and Korean students show no
difference in performance. At the first-year level, compared with the
Korean students, Chinese students present much lower percentage in
accuracy and higher percentage in wrong answer in the English way.
That may be correlated with their higher percentage in speaking English
outside classes than Korean students.

As in the previous tables, the vertical rows represent the subjects'
nationalities and levels. First, the results among American subjects is

presented in Table 10:

29

 

 

Table 10

Answers in Answers in total

Jap. way Eng. way
American 18 (72.096) 7 (28.096) 25
1st yr.
American 10 (71.496) 4 (28.5%) 14
2nd yr
American 28 (71.7%) 1105.396) 39
total

 

x2 -o.oo1 P-0.970>0.0S

As seen in part 1, about twenty five percent of wrong answers in English
way is found and the correct answers occupy about seventy percent. Chi
square shows that there is no significant difference between the first and
second-year levels.

Table 11 presents the comparison between American and Asian
students. The Asian subjects show a little better results than American

subjects, although the Chi square test shows that there is not a significant

 

 

 

difference.
Table 1 1

Answers in Answers in total

Jap. way Eng. way
American 28 (71.796) 11(25.3%) 39
total
Asian 24 (77.4%) 7 (22.5%) 31
total
X2 - 0.2 86 P - 0.593 > 0.05

In part 1, at the level, the percentage in "Answers in English way"

did not decrease in both Americans and Asians. In part 2, as for Americans,

30

the same result is found. However, in Table 12 of part 2, regarding Asians,
the accuracy rises in the second-year level; from 76.0 percent to 83.3
percent. With this improvement, there seen the decrease in percentage in
"Answers in English way"; from 24.0 percent to 16.6 percent. According to
the percentage in the table, there is a difference between the first and
second-year levels. (Chi square test is not applicable due to the small

numbers of the result.)

 

 

Table 12
Answers in Answers in total
Jap. way Eng. way
Asian 19 (76.0%) 6 (24.0%) 25
lst yr
Asian 5 (83.396) 1 (16.696) 6
2nd yr

 

(Chi-square test is not applicable)

Therefore, in the reading task, provided with ample time to think about the
answer, the second—year Asians performed better at answering negative
questions.

The first-year Asian subjects consist of Chinese and Koreans. But the
second-year subjects are only Koreans. When I focus only on Korean
subjects, it is found that there is no improvement observed in the second-
year level, and the percentage of "Answers in English way" is exactly the

same as in Table 13.

31

 

 

Table 13
Answers in Answers in total
139- way Big. way
Korean 15 (83.396) 3 (16.696) 18
lst yr
Korean 5 (83.396) 21(16.696) 6
2nd yr

 

(Chi-square test is not applicable)

Therefore, it is Chinese subjects who lower the performance of the first-

 

 

year level.
Table 14

Answers in Answers in total

Jap. way - Eng. way
Korean 15 (83.396) 3 (16.696) 18
lst yr.
Chinese 4 (57.196) 3 (42.896) 7
lst yr.
Asian 19 (76.096) 6 (24.096) 25
total

 

x2 - 1.895 P-0.169>0.05

The accuracy shown in "Answers in Japanese way" differs significantly
between Koreans and Chinese: 83.3 percent vs. 57.1 percent. Also, the
percentage of "Answers in English way" shows a big difference between
the two nationalities: 16.6 percent vs. 42.8 percent. In other words, Chinese
subjects performed much poorer than Korean subjects. However, since
there was no Chinese subject available in the second-year level, it was not
possible to investigate any change in Chinese students performance in the

second-year level.

32
Table 15 shows the percentage of speaking English, Japanese, and

their native language outside classes :

 

 

Table 15

96 of speaking

English Japanese native
language
Korean 28.8 96 7.2 96 63.3 96
lst yr.
Chinese 57.7 96 3.5 96 38.7 96
1st yr.

 

There is a big difference between Korean and Chinese subjects in terms of
speaking English and their native languages. Koreans use their native
language much more than English: Korean is spoken 63.3 percent, whereas
English is spoken 28.8 percent. Contrary to this ratio by Korean subjects,
Chinese subjects use English more than their native language: Chinese is
spoken 38.7 percent, while English is spoken 57.7 percent. It may be that
the high possibility that Chinese subjects are more influenced by English
since they are exposed to more English in their daily life. This could
explain why Chinese subjects have the higher percentage in "Answers in

English way" than Korean subjects.

When.

In this section, the subjects were asked to explain a schedule in
Japanese according to English cues. The English cues included four items
which contrast in word order in Japanese. The four items are arranged
horizontally in Table 16. Under each item, the subjects' answers are
categorized either in two ways, "Japanese way", (the right answer), or

"English way", (the incorrect answer).

33

duo—m Hm.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

muoo v! w :38 w: :05. M\N\HowA > 5853 a
.86. 52 man. .52 Eu. 28. mwmisme We. 2.8. 98:3. is. 5H E. 23
>3... 89.25 H H m A A m N m H m H H
H an 3.. mH.m8 HN.m8 uo8 88 Nm.m8 H H.A8 mm.H8 HA.N8
03:86 N N A A o A H w A A o A
H3 v4. mo8 8.8 H88 08 Nm8 Hm8 H88 08
88H m w H N m A H N u m H H Ho H H H
Hun kn. Hm8 Nm8 mmb8 ww.w8 NH.N8 H9H8 00.08 08
.33»: m H m m H m H m m m o m
Nada. 3&8 H908 3&8 Hm.m8 5&8 mw.w8 H88 08
8an HA A Hm Hm m H m A Hm HH Hm H H H
Hun an NEH HH.H8 NN.N8 HN.N8 NH.H8 Nw.m8 Hm.A8 oA.H8 m.m8
>328»: >828»: m m Hm w A Hm N H H Hw m A HN
H2 k... NH.H8 3&8 ow.N8 wo.H8 H m.w8 «A68 60.08 ww.w8
>328»: N m H m H .m o H H m H m
Naa v4. Nfim8 HH.A8 3&8 Hm.m8 o8 H008 3&8 368
SSH H Hw No HA m Hm N Hm No Hw m Hm
Hm" an Nan .38 $8 Hw.m8 Nm.w8 H08 008 HN.N8 NH.H8
8an NH HH mm NH Ho wH m wH wH No m um
mm.N8 AA.H8 HN.®8 NH8 Hm.N8 mw.H8 mN.m8 H H. H8

 

 

34
Table 17 compares the American and Asian subjects in their overall

performance, in which the four items are included. Asian students
generally did better than Americans: 32.8 percent for Asians and 53.2
percent for Americans. Although Asians' percentage of "Answers in
English way" is lower than Americans, it is still recognizable as
interference from English. However Koreans, the second-year students,

generally performed more accurately than the first-year students.

 

 

 

 

Table 17
J ap. way Eng. way total
Korean lst yr. 19 (63.396) 11 (36.696) 30
Exiifiéé’i'sl'; """"""" £17533}; """""""""" £13 136616
ELIai'iQE'yZ """"""""" éb'ié'siiBLSMWW"""1'613'Jf7'923""'"'"Ke’
REESE}? """""" HEBBEWWM""""7'1'2356?"""31
X5255}; """"""""" 5: ’7'1371186w'"""""'"’2'3’1’3'2'fs’éo'fmm’73
American 1st yr. 24 (47.096) 27 (52.9%) 51
XSQEEQJZAA'; """"" {2732.36'"""""""”122'5'3'342)"""""'2'c{
3.23.3132: """"""" 36mm)4113’3'f2'9'63"""""'7'7'
total 83 (56.496) 64 (43.596) 147

 

The Chinese students performed slightly better than Koreans. However,
depending on the items, different results obtained. For example, a poorer
performance was seen in the phrase AWL ("May 15,
1994") than in other items and in the item of W ("a
restaurant named Kyoto"), the students performed better than other items.
Therefore, for students, there is a phrase that has been acquired, while
there is another phrase, in which English word order is not still

differentiated from Japanese word order and transferred into Japanese.

35

Regarding the expression W! (5:00 I’M/AM) (see table
16), the incorrect answer in the English word order is found in higher
percentage among Americans. This influence from English does not
diminish in the second-year level and it even increases almost ten percent:
61.5 percent in the first-year level and 71.4 percent in the second-year
level. This may be due to the fact that first-year students learned this
expression more recently that second-year students, or due to the
dissimilarities between English and Japanese which causes the enduring
interference. However, compared with Americans, the Asian subjects
shows higher accuracy and a lower percentage in answers of "English
way": the average percentage in "English way" is 22.2 percent. Korean
subjects, in the second-year level, the percentage in answers of "English
way" increases: from 12.5 percent to 16.6 percent. Chinese subjects
present a much higher percentage in wrong answer than Korean subjects,
with fifty percent.

Another phrase of time expressions is W (" three
times an hour"). In this case, unlike the previous item, "5:00 PM",
American and Asian students show a similar result overall: about twenty
seven percent of "English way" is found in total among both Americans and
Koreans. However, Chinese subjects show no mistake this time. The first-
year Koreans present a high percentage in answers of "English way", that
is, fifty percent. In the case of both Americans and Koreans in the second-
year level, their accuracy rises and "English way" decreases: both
Americans and Koreans have 16.6 percent of "English way", which is
improved form 50.0 percent for Koreans and 30.7 percent for Americans.

In the expression WM ("May 15, 1994"),

"English way" is used much more often than "Japanese way". Regardless of

36
their levels, high percentages are marked by both Americans and Asians.

The students in their second-year level did even poorer than the students
in the first-year level in all nationalities. The second-year Americans
performed 100 percent in the "English way". Generally, Asian students did
better than Americans: 90.0 percent for Americans and 76.7 percent for
Asians. There is not much difference between Chinese and Koreans this
time. It seems that a general rule in Japanese that a bigger unit comes
before smaller units tends not to be retrieved and the English cue induces
the English order in which a small unit comes first.

Contrary to the previous result in W'May
15, 1994"), in an expression of W ("a restaurant named
Kyoto"), Asian students performed much better. The errors by Asian
subjects that are found in the first year disappear in the second year: 14.2
percent in the first year to 0 percent in the second year. Chinese students
did not make any mistake in this phrase. As for Americans, the first-year
students show twice as high percentage in answers of "English way" than
those of Koreans, a decrease to 16.6 percent in the second year. Therefore,
in this expression, the English influence is not seen as much as the other

expressions. Especially, Asian students show high accuracy.

W

In the fourth part, the students were asked to read sixteen sentences
in Japanese which consisted of four sentences in each four items. In each
item, there are two sentences that have the right word order and two that
have the wrong word order that reflects the English word order. The

subjects were instructed to mark their grammatical judgment on the four-

37

scale choices: "wrong", "probably wrong", "correct", or "probably correct".
In the following charts, if the right sentences were marked either "right"
or "probably right" and if the wrong sentences were marked either
"wrong" or " probably wrong", they are counted as "understanding". On the
other hand, if the wrong or right sentences were not identified as such,
especially if the subjects did not recognize the incorrect word order, those
sentences are counted as "not understanding". In "not understanding"
category, the subjects think that the English word order in the wrong
sentences is right or that the Japanese right word order in the right
sentences is wrong. Therefore, "not understanding" category is considered
to represent the English influence. Under " understanding" and "not
understanding" categories, there are sub-categories of "wrong sentence"
and "right sentence", which are, again, divided into two of four scales:
either "correct" and "probably correct", or "wrong" and "probably wrong".
Those categories are presented horizontally in the following charts.

The following Table 18 is the summary of part 4 that compares the
answers in the word order "Japanese way" and the word order in "English
way" among all nationalities and levels. In this part 4 Asians do not show
as high percentage of word order in the "English way" as Americans do.
One of the differences between the Asian students and American students is
that in the second-year level, the Korean students performed much better
than American students. Even in the higher level, the percentage of
answers in "English way" did not decrease much for American students.
The Asian students showed more certainty about their answers than the

American students.

38

 

 

 

 

Table 18

Jap. way Eng. way total
Korean 1st yr. 97 (71.396) 39 (28.6%) 136
EILEQEISI; """"""" I; 7653;; """""""" {8751,1333 """""" 252
232;}; yr """"""""" £563.38}; """""""" 5 2:729:53; """""" i'g's
{51325332}; """"""" 9'6'Eébii383mwm'""EZQEQQJW3'4'
X5513}; """"""""" 556293.696"""'7 """" E. 32513333""""""'2§2
American lst yr. 114 (59.0%) 79 (40.996) 193
$583533; """"" 92:72.38} """"""""" A ’273’536"""""""1'zb
$555135: """"""" i5'il$'{.3'9},3"""""""153135266;""""""31'3
total I 412 (69.296) 183 (30.7%) 595

 

Table 19 presents the result of a item, W ("5:00 PM"). Among
all subjects, 36.3 percent of "not understanding" and 63.6 percent of
"understanding" are found. 44.2 percent of the Americans are in the
category "not understanding" in all levels, and show an improvement in
their second-year level. Asian students show a still lower percentage in
"not understanding", 27.4 percent, than Americans, and in their second-
year level, they also show a lower percentage than the percentage in their
first-year level: from 30.6 percent to 15.3 percent. As for Chinese students,

their percentage is about 10 percent higher than that of Koreans.

Table 1

sentence

39

SL‘I‘IICHCC

choice

1 st
Chinese
1 st .
total

1 st
Korean

2nd
total
st 8: 2nd

l st
American
2nd .
total

8:

COITCCI SOHlCflCC

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

not understanding (ling. way) tot-.tl
kind of sentence wrong sentence correct sentence
subject's choice correct probably wron probably
correct wrong
Asian Korean 4 2 6 3 3 9 33
1st yr. 18.196 . 9% 27.2%
Chinese 3 0 3 0 3 6 16
1st yr. 18.7% 18.796 37.596
total 7 2 9 3 6 15 49
lst yr. 18.396 12.296 30.6%
Korean I 0 l l l 2 13
2nd yr. 7.6% 7.696 15.396
total 8 2 10 4 7 17 62
lst & 2nd 16.196 11.296 27.4%
American American 8 7 IS 3 7 22 46
lst yr. 32.6% 15.2% 47.8%
American 4 4 8 0 I 9 24
2nd yr. 33.396 4.196 37.5%
total 12 1 l 23 3 8 31 70
1st & 2nd 32.896 11.496 44.296
total 20 13 33 7 15 48 132
2596 1 1.396 36.396

 

 

 

 

 

 

 

 

 

 

 

40
Table 20 below shows the subjects' certainty about their answer:

whether they choose "right" and "wrong", or "probably right" and
"probably wrong".

Table 20

 

sure unsure total.
"right" or " wrong" "probably _"

 

Americans 37 (52.896) 33 (47.196) 70
Asians 42 (67.796) 20 (32.296) 62
total 79 (59.996) 53 (40.196) 132

 

x2 - 3.031 P-0.082>0.05

The Americans were more uncertain about their answers than Asians. In
general, about forty percent of uncertainty obtains with regard to the word
order of gQgQ_gg_-jj ("5:00 PM"). Almost half of Americans express their
uncertainty.

Table 21 concerns the phrase WW (" May 15,
1994"). In this table, a similar tendency to the previous "5:00 PM" is found.
In general however, among the Asians, this "May 15, 1994" phrase obtained
poorer results than the previous phrase, 8981892116530 PM"). This is true
especially of the case for the first-year Asians. Also, an improvement in
performance from the first-year to the second-year level is found. Again,

the Chinese subjects show poorer results than the Korean subjects.

Table 2] Q/ 1 5/19941

41

 

understanding (lap. wa L

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

wron sentence correct sentence
wrong probably correct probably
wrong .3 correct __ ,
Asian Korean 5 2 7 14 3 17 24
lst yr. 2096 48.596 68.596
Chinese 1 1 2 7 0 7 9
1st yr. 12.596 43.796 56.296
total 6 3 9 21 3 24 33
lst yr. 17.696 4796 64.796
Korean 6 4 10 10 0 10 20
2nd yr. 43 .496 43.496 86.996
total 12 7 19 31 3 34 53
lst & 2nd 25.696 45.996 71.6%
American American 5 2 7 13 6 19 26
lst yr. 1496 25.696 35.196
American 4 l 5 9 2 l l 16
2nd yr. 18.596 40.796 59.296
total 9 3 12 22 8 30 42
1st & 2nd 15.596 38.996 54.596
total 21 10 31 S3 11 64 95
20.596 42.396 62.996
not understandin lin . wily) total
kind of sentence wrong sentence correct sentence
subject’s choice correct probably wrong probably
correct wroi
Asian Korean 5 6 1 1 0 0 0 l l 35
1st yr. 31.496 096 31.496
Chinese 4 2 6 0 1 1 7 16
1 st yr. 37.596 6.296 43.796
total 9 8 I7 0 l l 18 51
lst yr. 33.396 1.996 35.296
Korean 1 1 2 0 1 l 3 23
2nd yr. 3.996 4.396 13.96
total 10 9 l9 0 2 2 21 74
lst & 2nd 25.696 2.796 28.396
American American 13 S 18 5 l 6 24 50
1 st yr. 36.96 I 2.96 32.496
American 7 2 9 l 1 2 I l 27
2nd yr. 33.396 7.496 40.796
total 20 7 27 6 2 8 35 77
lst a 2nd 35.96 10.396 45.496
total 30 16 46 6 4 10 56 151
30.496 6.696 37.96

 

 

 

 

 

 

 

 

 

 

 

42

However, with regard to the subjects' certainty about their answers, the

item W (" May 15, 1994") were higher percentages

of certainty than the previous item. Almost 73 percent of sureness is

attained in this item, whereas sixty percent in W ("5:00 PM"). This
time, Americans show higher certainty than Asians, although Americans

have lower accuracy than Asians.

 

 

Table 22
sure unsure total
"right" or " wrong" "probably _"
Americans 57 (74.096) 20 (25.996) 77
Asians 53 (7 1.696) 2 1 (28.396) 74
total 110 (72.896) 41 (27.196) 151

 

Table 23 in the next page shows the result with W
("three times an hour"). Overall, the high percentage of "understanding",
is shown at 77.8 percent, while 22.1 percent of " not understanding"
obtains. The Asians did much better than the Americans: 13.3 percent was
categorized in "not understanding" by the Asians compared with 31 percent
by Americans. Among both the Americans and the Asians, the second-year
subjects show better results than the first-year subjects. Koreans
especially, show a big improvement, which is from 19.4 percent to 8.3
percent. Among the first-year subjects, the Chinese subjects performed
more accurately than the Korean subjects: only 6.6 percent in "not

understanding".

Table 23 (3 times an hour

43

 

understanding (Jap. wa ')

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

wrong sentence correct sentence
wron probably correct probably
wronj correct
Asian Korean 6 7 13 12 4 16 29
lst yr. 36.196 44.496 80.596
Chinese 4 2 6 5 3 8 14
lst yr. 4096 53.396 93.396
total 10 9 19 17 7 24 43
lst yr. 37.296 4796 84.396
Korean 7 3 10 ll 1 12 22
2nd yr. 41.696 5096 91.696
total 17 12 29 28 8 36 65
1st & 2nd 38.696 4896 86.6%
American American 6 8 14 9 9 18 '32
1 st yr. 23.796 30.596 54.296
American 7 2 9 8 2 10 19
2nd yr. 3696 4096 7696
total 13 10 23 17 1 1 28 51
lst & 2nd 3196 37.896 68.996
total 30 22 52 45 I9 64 l 15
34.896 42.996 77.896
not understandinmg. way) total
kind of sentence wrong sentence correct sentence
subject’s choice correct probably wroll probably
correct wrong
Asian Korean 4 l 5 2 0 2 7 36
lst yr. 13.896 5.596 19.496
Chinese 0 l 1 (i (l i) 1 15
lst yr. 6.696 096 6.696
total 4 2 6 2 0 2 8 51
lst yr. 1 1.796 3.996 15.696
Korean 2 0 2 0 0 (i 2 24
2nd yr. 8.396 (196 8.396
total 6 2 8 2 0 2 ll) 75
lst & 2nd 10.696 2.696 13.396
American American 4 5 9 2 6 8 17 59
lst yr. 15.296 13.596 28.896
American 2 3 5 0 l l 6 25
2nd yr. 2096 496 2496
total 6 8 14 2 7 9 23 74
1st 8 2nd 18.996 12.196 3196
total 12 10 22 4 7 11 33 149
14.796 7.396 22.196

 

 

 

 

 

 

 

 

 

 

 

 

44

As for the subjects' certainty about their answers, a big difference
between the Asians and the Americans is found. The Chi square test shows
the difference. The Asians are more sure of their answers than the

Americans: in "understanding", 70.6 percent by Asians and 51.3 percent by

 

 

Americans.
Table 24
sure unsure total
"right" or " wrong" "probably _"
Americans 38 (51.396) 36 (48.696) 70
Asians 53 (70.696) 22 (29.396) 62
total 91 (61.096) 58 (38.996) 149

 

x2 -5.845 P - 0.016 <o.os

The result with the last item, Wave restaurant
named Kyoto"), is shown in Table 25. In this phrase, the first-year Chinese
subjects show high accuracy with only 6.6 percent of "not understanding",
whereas the first-year Koreans show 37.5 percent, which is slightly higher
than that of the first- and second-year Americans. However, this 37.5
percent by the first-year Koreans decreased about by 30 percent in the
second-year level. As for Americans, the second-year students do not
improve and they even show higher percentage in "not understanding":
33.3 percent in the first-year level and 36.3 percent in the second-year

level.

45

Table 25 (a restaurant named Kyoto)

 

understanding (Jap. wa )

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

kind of sentence wrongsentence correct sentence
subject's choice wrorg probably correct probably
wron correct
Asian Korean 5 2 7 9 4 13 20
1st yr. 21.896 ’ 40.696 62.596
Chinese 6 0 6 6 2 8 I4
1 st yr. 496 53.396 93.396
total 11 2 13 15 6 21 34
lst yr. 27.696 44.696 72.396
Korean 12 0 12 1 1 0 l 1 23
2nd yr. 5096 45.896 95.896
total 23 2 25 26 6 32 57
1st 8: 2nd 35.296 4596 80.296
American American 8 4 12 15 5 21) 32
lst yr. 2596 41.696 66.696
American 5 3 8 15 5 20 28
2nd yr. 18.196 45.696 63.696
total 13 7 20 30 10 40 60
1st & 2nd 21.796 43.496 65.296
total 36 9 45 52 12 72 117
27.696 44.196 71.796

 

sentence
choice

Korean
lst

1 st

total

lst
Korean
2nd

total

lst & 2nd

lst

2nd
total
st

SCfllchC

COITCC1 sentence

 

 

46
In this item Wm("a restaurant named Kyoto"), the

Asian students feel most confident about their answers: 81.6 percent of
sureness obtains. Also, the Americans show relatively high certainty, 72
percent, although their accuracy is not high, which is 34.7 percent of

incorrect answers, which is shown in the word order test.

 

 

Table 26
sure unsure total
"right" or " wrong" "probably _"
Americans 54 (7296) 21 (2896) 75
Asians 58 (81.696) 13 (18.396) 71
total 112 (76.796) 34 (23.296) 146

 

x2 .. 1.7=917 P-0.166> 0.05

The certainty of all four items is presented in Table 27:

 

 

Table 27
sure unsure total
"right" or " wrong" "probably _"
Americans 186 (63.696) 110 (37.296) 292
Asians 206 (76.596) 76 (28.296) 269
total 392 (67.896) 186 (32.196) 578

 

x2 - 6.900 P - 0.009 <0.os
The Chi square test shows that the results by the Americans and the Asians
are different. It shows that the Asians feel more confident with their

answers and the word order.

47

.H.»Zn Nm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Too :3 w :38 »= :9: m\N\HooA > :»3nn w
3.2» m:.<<» lll_»m.<<»N m:.<<» .11.».5» m:.<<» lhmmefik m:.$.» 1.1
>43»: .33»: NA m w& No H mm. NA H H &m No HN wN
Hun k... HN.H8 NH.N8 mo.m8 Ho.A8 om.m8 wH.A8 mN.m8 wH.m8
03:3» Ho 0 Hm HA H Hm o H Hm HA H Hm
Hm" k... mN.m8 &H&8 3&8 m.m8 mm.N8 Aw.H8 3&8 0.08
SE &A Hm Aw Aw m mH && Hm mH wA Hm AH
Hun v4. mo&8 wo.o8 wAurw8 Hm.m8 0A.H8 wm.N8 HN&8 NH.m8
#03»: HH N Hm NN N NA No & N& N& H NA
ime 3.. @908 5&8 268 m&8 mmb8 H&8 8&8 A. H8
88.. Am HH mN 3 Ho Hm mm. NH HA MH HA HH
Hm" m" N:a HN.m8 NH.A8 «908 H&&8 1 HH.m8 Nm&8 mmPN8 HPH8
>328»: >3m1n»: NA NN Am wN HH mm Nm NA mo wN Hm Am
Hun v4. wNH8 AH.m8 mA.N8 Nm&8 me8 wN.A8 8&8 ww&8
>323»: Hm 0 NA Ho 0 Nu Hm HH NH Nm Hm AA
Nan— Vq. 8&8 wH.m8 Hm8 NA8 mm.N8 Ao.H8 1mmb8 wmum8
88— we wH Ho H N& HA AN wm HH mo wN wN
Hm" an NEH mm.H8 AA.N8 0908 wH8 mA.m8 Am.A8 3&8 &A.H8 1.
82 mA Am HwN HHm && HAO mm mm. HwH HHH Am How
338 &m&8 HH.m8 NN.H8 mNb8 wH8 HH.H8 1Nlm.N8

 

 

 

 

 

48

Table 28 is the summary of results including all four items.
Comparing four items together, it is clear that the first-year Koreans and
Americans show a relatively high percentage in "not understanding
(English way)". The results vary depending on items, which can be
explained by some reasons: it can be due to the time when the students.
learned the item (The later, the more fresh their memory is.) and the
frequency of the usage. In the second year, the Koreans become more
accurate, whereas Americans do not show much improvement in their
second year. The Chinese subjects sometimes do well and sometimes do
poorly: high percentage of "not understanding (English way)" is seen in
"5:00 PM" and "May 15, 1994" items, but not in the other items.

Let us compare Table 28 in the previous page in part 4 to Table 16 ,
which is the result of the word order test of oral test in part 3, to see the
difference between the oral and written contexts. First, generally the
result in part 3, the oral context, has higher percentages in "English way"
than the result in part 4. As for the degree of improvement from the first-
year to the second-year, it is more apparent in part 4 than in part 3. The
second-year Americans in part 3 show even a higher percentage in the
English way than the first-year Americans. The improvement seen in the
Koreans in part 3, which is from 36.6 percent to 29.1 percent, is not as
much as the improvement in part 4, which is from 28.6 percent to 9.5
percent. The Chinese subjects also show a better result in part 4 than in
part 3. Thus, generally, the written context results in better performance,
which exhibits less of the "English way" than in oral test.

Table 29 is an overall summary of all parts that includes part 1 to
part 4. It compares the final ratio of "Japanese way" and "English way",

and Americans and Koreans. The data collected from Chinese subjects is

49
eliminated from the table since their data tend to be a little different from

the data from the first-year Koreans. The general conclusion is that the
answers that reflect English features are found among Asians as seen in
Americans although the percentage in Asians is lower than that in
Americans. Both Americans and Asians perform better in the written
context than in oral context. In negative questions task, in Part 1 and 2, no
decrease of percentage in the "English way" in the second year is seen. For
negative question in the oral test, it even increases and at the same time,
the avoidance strategy is made use of in the second year, too. As for word
order task, in Part 3 and 4, however, answers in the "English way" decrease
in the second year, especially in the written context. Therefore, there is no
consistent result through all parts, nationalities, and levels, and the results

really vary depending on each variable.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Table 29
Part 1 Part 2
Jap. way Eng. way Jap. way Eng. way ,
Korean lst yr. 49.3 96 21.6 96 83.3 96 16.6 96
Korean 2nd yr. 34.3 96 28.3 96 83.3 96 16.6 96
Korean total 41.8 96 24.9 96 83.3 96 16.6 96
American 1st yr. 57.8 96 26.4 96 72 96 28 96
American 2nd yr. 28.3 96 23.8 96 71.4 96 28.5 96
American total 47.3 96 25.5 96 71.7 96 25.3 96
total 42.5 96 24.8 96 75.6 96 24.2 96
Part 3 Part 4
Jap. way Eng. way Jap. way Eng. way
Korean 1st yr. 63.3 96 36.6 96 71.3 96 28.6 96
Korean 2nd yr. 70.8 96 29.1 96 90.4 96 9.5 96
Korean total 67 96 32.8 96 80.8 96 19 96
American 1st yr. 47 96 52.9 96 59 96 40.9 96
American 2nd yr. 46.1 96 53.8 96 65 96 35 96
American total 46.7 96 53.2 96 61.3 96 38.6 96
total 56.4 96 43.5 96 69.2 96 30.7 96

 

 

 

 

 

 

 

50
W

The present study investigates the 1.2 interference in L3,
particularly in learning negative questions and word order of certain
phrases. The results of the data indicate that the Asian students learning
Japanese as a third language have some interference from English, though
to a lesser degree than the L1 interference for the Americans. In the study
by George (1972), one third of errors was due to transfer. In this study,
generally, about 28.5 percent of English influence in oral context and 21.1
percent in written context are found in Asians.

However, there remained a problem: it has to be proven that the
subjects are really influenced by English and they are transferring it into
Japanese. The subjects' mistakes may be due to English interference, but
they may also be due to developmental problem in learning a new
language. How can we tell that transfer is really happening? in chapter
two, I discussed the ambiguous relationship between transfer and
developmental errors, which is not a clear-cut but an interactive
phenomenon. As Zobl (1982) states, it is hard to differentiate the two of
them.

Major (1987) states that interference processes are predominant at
the early stage of acquisition and they decrease over time, while
developmental processes increase. In this study, although there are only
two levels to compare, Major's theory applies to the word order test in Part
3 and 4. The number of errors caused by English structures diminish at the
higher level. However, his theory does not apply to the case of negative
questions in Part 1 and 2. Even in their second-year level, English way of
answering does not decrease. However, to claim this theory, further

longitudinal study that includes more levels should be conducted.

51
According to the present study, the transfer from English system in

Japanese negative questions does not decrease much even in the second
year. First, the reason is that originally, perhaps to answer English
negative questions is easier than Japanese negative questions. in English
answers, "Yes" precedes the affirmative, "I do", and "No" precedes the .
negative, "I don't" However, to answer Japanese negative questions, "Yes"
means negative and "No" means affirmative. In addition, an answer in
Japanese has to do with both the questioner's presupposition about the
matter involved, whereas in English, an answer depends on an answerer's
intention about the matter. Therefore, grammatically and pragmatically,
English system of negative questions tends to be simpler than Japanese
system.

Secondly, for the Asian subjects, we can speculate that their
conscious efforts to acquire and use the answers in the English way forces
an automatic response in the English way even when asked a question in
Japanese. It may be possible that for Asians who start living in an English
speaking environment, there is a tendency that it takes time and efforts to
acquire the way of answering negative questions in English, which was
the opinions given by the subjects in this study. When they try to answer,
they are conscious not to use the way of their native language but to
respond in English way. It is possible that this conscious efforts in
speaking English function unconsciously in Japanese, too, even if they are
taught the Japanese negative questions.

It was discussed in chapter two that dissimilarities between 1.1 and L2
cause slower development (e.g. Keller-Cohen, 1979, Schumann,l982). In
the present, it is found that sometimes the dissimilarities between 1.2 and L3

(target language) cause the transfer from 1.2 to L3, which could stay at a

52
higher level even when there are similarities between 1.1 and L3. In

negative questions, it is possible that the dissimilarity between 1.2 and L3
or 11 and L2 is so outstanding that the advantage of similarity between 1.1
and L3 is erased. This could be a disadvantage for multilinguals. If one
knows several languages, there may be several competing different '
parameters in a learner's head and one may chose wrong parameters in
learning a new language. Therefore, in this case, knowing a language that
relates to 1.3 closely may not be advantage. This study shows that the
likelihood of transfer lies in the situation when there is some distance
between previously learned languages.

Additionally, the dissimilarity between L2 and L3 and its resultant
difficulty for students brings about avoidance of the structure as Schachter
(1974) mentioned. The present thesis shows one example of this avoidance.
In part 1, negative questions in conversation, about one third to half of
answers lack "Yes/No". The interesting thing is that the ratio of this
avoidance is higher among the second-year level American and Asian
students. However, the answers without saying Yes/No is not incorrect but
common in native speaker's speech. The avoidance found in this study is
different in this point from Schachter's avoidance. However, the
avoidance in this study still shows a learners' strategies and the influence
from the previously known languages.

In addition to this avoidance, some self-correction from "English
way" to "Japanese way" is also found, although the percentage is not very
high and it varies from five percent to thirteen percent depending on
nationalities and levels. According to Krashen (1983), self-correction and
avoidance can be accounted for by his monitor theory. The monitor refers

to the efforts in conscious language learning to control one's language

53

production. The results in this thesis that sometimes the subjects were
making use of not saying Yes/No and self-correcting their errors may be
accounted foe by his theory, too. The subjects were sometimes conscious
not to make errors which is caused by the transfer from English. The data
in this study is similar to his theory. It is possible that Krashen's monitor
theory can also apply to multilingual learning situations like this study.

One of the general results is that the subjects perform better in the
written context than in the oral context. However, that does not mean that
there is no influence from English in written text of Part 2 and 4. Ulijin
and Kemmpem (1976) state: errors caused by 1.1 transfer should not be
found in understanding of written text. Although the higher accuracy is
found in the reading test than in the oral test, there is still some English
transfer in their answers in the reading test. The subjects sometimes
cannot find the ungrammatical structures in Japanese that reflect English
patters in reading situation, though in reading test the subjects can have
enough time to pay careful attention to the structures than they can in oral
test. '

The reason why the higher accuracy is found in the written context
lies in the difference between spoken and written discourses. In casual
spoken discourse, the speaker pays more attention to content than to form
in a situation where a prompt response is required. 0n the other hand, in
written discourse, the reader has more time and careful attention to both
content and form. Therefore, it is natural that the test in written context
shows higher accuracy than the test in conversation.

Lastly, the question that Magiste (1979, 1986) and Balke—Aurel and
Lindblad (1982) try to answer is: is a third language easier for bilingual

than learning a second language is for monolinguals? In the present

54

paper, overall, bilingual Asians, do better than monolingual Americans.

In this study's case, the fact that L1 and L3 are close for Asians may help in
many ways. In this sense, it is not advantageous for American whose L1
has many dissimilarities to learn Japanese as L2. Therefore, generally
speaking, learning Japanese as a third language is easier for Asians than it
is as a second language for Americans because of many similarities.
However, it cannot be neglected that the bilingualism in Asians may have a
negative impact on their L3 learning, too. Knowing and actively using L2
gives rise to the transfer from L2 to L3, blocking the help from similar
parameters in L1 to L3. It has to be acknowledged that it is not always easy
for bilinguals to learn L3 and it is likely that the bilingualism could result
in difflculties or some errors which are caused by knowing a few

languages, although this depends on what L1, L2 and L3 are.

m

The present study investigates bilinguals learning L3. The Korean
and Chinese subjects, who are bilinguals in their native language and
English and who are learning Japanese as L3, are compared with
Americans studying Japanese as L2 at Michigan State University. There are
nineteen Asians and twenty Americans as subjects who are chosen from
Japanese first and second-year levels. Negative questions and some
phrases in Japanese which contrast in word order in English are selected
as the items for the tests. The main questions were the existence of 1.2
transfer into L3 and a question: does bilingualism have any impact on an
individual's L3 learning? . The results between two levels and two contexts,

conversation and reading, are compared.

55

In my experiment, it is found that Asian students make the same
mistakes that Americans make. Those mistakes in Japanese reflect English
grammar and structures, which indicates the English transfer into
Japanese both in Asians and Americans. Although the percentage of the
English influence varies depending on nationalities, levels, and chosen
items, Asians show higher accuracy in general and higher accuracy in
their second year than Americans. The interference of English in
Japanese is more persistent in monolingual Americans. The word order test
presents '-igher percentage of the influence from English than the
negative question test. However, the results in negative question test
reveal persistent errors caused by the English system even in the second
year in both nationalities. In the second—year level, more subjects begin to
make use of "monitor" to self-correct errors and to avoid saying geijg
("Yes/No . hat is different from the English system. In some cases, the
similarity bet "een L1 and L3 in Asians is of no use and may be erased by
the dissimilarity between 1.2 and L3. The students attain higher accuracy
in reading context than in oral context and less influence from English is
seen In the written context. In a prompt response, the transfer more likely
occurs.

'lhese results indicate that when learning Japanese as .i third
language, circumstances favor bilingual Asians. Howewr, the
recognizable nt- :atbers of errors caused by the interference between L2 and
L3 suggest that there may be competing systems of previously learnt-«.1
languages in the learners' cognitive system. Sometimes a different
parameter from the languages previously learned may be rctr‘wed. These
can give us some insight about how learners make some en ors. In

addition. pedagogically, the finding in this study reminds us that when

56

Korean and Chinese learn Japanese (the case that L1 is close to L3), it is
sometimes helpful to mention the similarity between L1 and L3 if possible.
It is not necessarily an easy task for adult bilinguals or multilinguals to
learn a new foreign language.

The present paper points out one of the possible areas where
multilingualism works disadvantageously. However, the study of
multilingual acquisition is usually too complicated and diversified to reach
a general conclusion. Further study on multilingual acquisition from
different angles is necessary in order to attain better grasp of language

learning in general.

Appendix I
Part 2
Please read the passage and answer the questions.

Mliékh/ufiffifztflikmfiéibt. assumflfiéitbiaftluf‘d'. waéfit<3lv8sa
'CL‘tflI‘. abatsescmmuestthrbm é’luhlu't‘bk.

(1) coaMiat'x—A‘Cfiei 61:1».
1 xx.fi2$bt. 2 uui.fi$$thrbt.

(2) CwAliflllKTésUmNfiéfiibtb‘.
1 ii. fiétbt. 2 Ui‘ft. fiéifih‘t’bt.

(3) ’iUfllNfiék’X amok/111'???»

1 ii. fiétb‘ot’t’i‘. 2 ii. fi$t<bU$fitht.

3 “Hi. fiéttfiat‘t’i’. 4 uwt. fi8t<bU$filvat.
(4) fifiS‘banUfllNfiéifl/V‘CLIUJ‘.

1 ii. fiéibt. 2 ii. fi‘é’it’h‘t’l’k.

3 unit. 1533111122. 4 Milk. fiéifi/u‘t’bt.

Part 4

Please read the following sentences. Then, if you are sure that a sentence is
ungrammatical, please circle 1, if it is probably ungrammatical, circle 2, if it is
probably correct, circle 3, if you are sure that it is correct, circle 4. Also,
please underline the places where you think they are wrong or probably
wrong.

wrong probably probably

correct
wrong correct

1 Cwbeb/uii1990$5§181€t9$b7§3 1 2 3 4
2 N m4 4 mFi/t‘ct. t 2 3 4
3 5>9>dtt\5t:61:1'h'ct\$1'. l 2 3 4
4 xuyewatita4ztzauar. 1 2 3 4
5 4mm afitzrhbbi5926t. 1 2 3 4
6 ilskt: 157/ thfifttfifiéfiz we. 1 2 3 4
7 seetxtsanemzna 1.1:. 1 2 3 4
8 e"v‘lfllatllMElt992i‘li‘t'1”. l 2 3 4
9 5 El 1 ilmtcxz‘lmfiéit. 1 2 3 4
10 aoasoxtuameehtzauzbm 1 2 3 4
11 thbxiataaflsatsssifi‘ct. I 2 3 4
12 1 filcaillllfibb‘bvit. l 2 3 4
13 mwetttiiBMtc-Dtb'tuit. 1 2 3 4
14 4.309951131816113 859:1. 1 2 3 4
t 5 717.6114th maintain 1 2 3 4
16 “1319882585813192/ubibt. 1 2 3 4

57

Appendix II

Part 5

Answers to the following questions regarding your Japanese study in the U.S.
will be used as a part of the data in my project only if you give me verbal
permission to do so. At the same time, no names are attached to the data.
Pleases answer the following questions.

- the length of your stay in the U.S. and the length of your study of English

 

- your TOEFL score and the date when it was taken (if you don't mind)

 

- your level of Japanese first-year second-year

- What percentage of time do you spend speaking outside of your classes?

1. English- 96 2. Japanese - 96 3. your native language-- 96

 

- Do you have any difficulty in using English in daily life, including
studying?

 

- When you study Japanese, do you usually use (or think in) your native
language?

 

- Do you think that studying Japanese in English is inefficient, difficult, or
uncomfortable?

 

- Do you wish you could study Japanese in your native language?

 

Do you think that studying Japanese is easy in general?

 

58

BIBLIOGRAPHY

Akiyama, M. Michael. 1979. Yes-No Answering Systems in Young Children.
in Cognitive Psychology 11. 485-504.

Bailey, Nathalie. Carolyn Madden. and Stephen D Krashen. 1974.15 there a
"natural sequence" in adult second learning" language Learning
24: 2. 235-243

Balke-Aurell, G. & Lindblad, T. 1982. Immigrant students and their
language. Report No. 23. Dept. of Education Research. University
of Gothenburg, Sweden. 79-89.

Cenoz, Jasone and Jose F. Valencia. 1994. Additive trilingualism: Evidence
from the Basque Country. in Applied Psycholinguistics 15. 197-
209.

Chamot, Anna U. 1978. Grammatical Problems in Learning English as a
Third language. in Marcussen, E. Hatch (ed.) second Language
Acquisition: A Book of Readings. 175-190.

DiPietro, R. 1961. Borrowing: its effect as a mechanism of linguistic
change in American Sicilian. General Linguistics. 5. 30—36.

Dulay, Heidi C. and Marina K. Burt. 1974 a. Natural Sequences in Second
language Acquisition. language Learning 24:1. 37-53.

Gass, Susan. 1979. language Transfer and Universal Grammatical
Relations. Language Learning. 29: 2. 327-344.

Gass, Susan and Larry Selinker. 1983. Introduction. in Cass and Selinker
ed. 1983. language transfer in language learning. Mass.
Newbury House Publishers. 1-18.

Gass, Susan. 1985. A Review of lnterlanguage syntax: Language transfer
and Language Universals. Language Learning. 34: 115-132.

Ilome, Elinor Clark and Yun, Sang Soon. 1950. Introduction to spoken
Korean Volume I. Connecticut: Yale University.

Jaspaert, Koen and Gertrud. 1989. linguistic Evaluation of Dutch as a Third
Language. in Byram, Michael and Johan Leman (eds.), Bicultural
and Trilingual Education: The Foyer Model in Brussels. Clevedon:
Multilingual Matters Ltd.

Keller-Cohen, Deborah. 1979. Systematically and variation in the non-
native child's acquisition of conversational skills Language
Learning. 29: 27-44.

Klein, Wolfgang. 1986. Second language acquisition. New York: Cambridge
University Press. 23-32.

Kovac, D. 1965. On psychological problems of commanding more languages
by an individual. Studia Psychilogica. 7. 158-159.

59

60

Krashen, D. Stephen. 1978. Individual variation in the use of the Monitor.
in W. Ritchie ed. Principles of second language learning. New
York. Academic Press. 175-183.

Krashen, D. Stephen. 1983. Newmarkk's "ignorance hypothesis" and
current second language acquisition theory. in Gass and
Selinker. 1983. Language transfer in language learning. 135-153.

Kuno, Susumu. 1973. The structure of the Japanese Language.
Massachusetts: The MIT Press.

lado, R. 1957. Linguistics across Cultures. Ann Arbor. University of
Michigan Press.

Larsen-Freeman, Diane and Michael H. Long. 1991. An introduction to
second language acquisition research. New York. Longman. 1-9.

Li, Charles N. and Sandra A. Thompson, 1940. Mandarin Chinese, A
Functional Grammar. California: University of California Press.

Mackey F. William. 1968. the description of bilingualism. in Fishman, J. A.
ed. Readings in the Sociology of Language. The Hague. Monton.
554584.

Magiste, Edith. 1986. Selected Issue in Second and Third Learning. it
Language Processing in Bilinguals: Psycholinguistic and
Neuropsychological Perspectives. ed.. Jyotsna Vaid. lawrence
Erlbaum Associates Publishers. London. 97-122.

Major, Roy C. 1987. A Model for lnterlanguage Phonology. in Ioup,
Georgette and Steven H. Weinberger (eds.) Interlanugage
Phonology: The Acquisition of a Second Language Sound System.
Cambridge, MA.: Newbury House Publishers. 101-124.

Mclaughlin, Barry. 1984, 1985 (First ed. 1978). Second-language
acquisition in childhood, 2nd ed.. 2 vol. Hillsdale, NJ: lawrence
Erlbum. 59-71, 215-231.

Odlin, Terence. 1989. Language Transfer: Cross-linguistic influence in
language learning. New York: Cambridge University Press.

Ree, Jung-No. 1974. T0pics in Korean Syntax with Notes to Japanese. Korea:
Yonsei University Press.

Richards, Jack C. (ed.) 1974. Error Analysis. London: Longman.

Schachter, Jacquelyn. 1974. An error in error analysis. Language
Learning. 24: 2. 205- 214.

Schumann, II. John. 1979. The acquisition of English negation by speakers
of Spanish: a review of the literature. In R. Anderson ed.. The
Acquisition and Use of Spanish and English as First and Second
Languages. Washington, D.C.:TESOL. 5-32.

61

Schumann, J. 1982. Simplification, transfer and relexification as aspects of
pidginization and early second language acquisition. language
Learning, 32

Selinker, Larry. 1966. A psycholinguistic study of language transfer.
Unpublished Ph.D. dissertation, Georgetown University.

Taylor, 8. P. 1975. The use of overgeneralization and transfer learning .
strategies by elementary and intermediate students of ESI. language
Learning, 25: 73-107.

Ulijin, J. and G. Kempen. 1976. the role of the first language in second
language reading comprehension-some experimental evidence.
Proceedings of the fourth international congress of applied
linguistics. Stuugart. 495-507.

Wagner, Daniel A. and Jennifer E. Spartt. 1989. Does learning to read in a
second language always put the child at a disadvantage? Some
counter evidence from Morocco. In Applied Psycholinguistics. 10. 31-
48.

Zobl, Helmut. 1982. A direction for contrastive analysis: the comparative
study of developmental sequences. TESOL (htarterly, 16: 169-184.

       

     

    

    

     

 

HICH N STRT

IGQ
IHIWIIIHIIIII |
3129

3

UNV.L

t I 1111112111111“

0445

 

 

 

 

 

 

 

 

E
"MINI
01