chapter i introduction - web viewchapter one. 1.1 . introduction. the closer the world gets...
TRANSCRIPT
Chapter One
1.1 Introduction
The closer the world gets together, the more there is
a need for communication with people who speak different
languages. Speaking second or third languages fluently is
one of the biggest desires for those who wish to succeed.
Furthermore, the fact that one can use a second language
will be very common among people. Therefore, the
importance of investigating second language acquisition
(SLA) is clearly supported given this tide, and therefore,
scholars have developed various areas of research in SLA
and as a result, this study has become a central research
area of linguistics since the 1970’s. Even though they
adopted various ways to investigate how people acquire
second languages or what makes it possible to learn second
language efficiently, their core questions were similar. A
central theme has been “Is there any sequence in order of
acquisition of certain languages?” or “Why do some language
features emerge earlier than others?” (Hatch)1.
Among the various fields that researchers have
examined include studies on developmental sequences.
Throughout these studies, developmental sequences can be
defined as “the order in which certain features of a
language…are acquired in language learning” (Lightbown and
Spada)2. These are very interesting to look at because it
seems to be almost impossible to define a general image of
developmental sequences in second language acquisition in
spite of its clear concept.
In the following section, several key words will be
explained to clarify the idea of developmental sequences in
second language acquisition: These questions are
1) what do developmental sequences mean? and,
2) where can researchers find developmental patterns?
1.2 Developmental Sequences
Developmental sequences always occur when people
learn any kind of language as first or second. In first
language acquisition, which generally happens at the early
stage in childhood, there can be seen a general
developmental pattern for certain grammatical elements such
as the present progressive, plurals; irregular past;
possessive “’s”; copula; articles; regular past; 3rd person
singular; simple present auxiliary and so on.
Children acquire their native language using several
methods, not only by imitating adults but also experiencing
and discovering the structures of their own L1. It is very
interesting that almost all children pass through certain
stages similarly in L1 acquisition in spite of their
- 2 -
different environment. For example, for grammatical
features, most of the children will acquire present
progressive -ing earlier than the copula, and plural -s is
earlier than possessive ’s. In negation, first, they start
using negatives alone or as the first word in the
utterance, then gradually, they combine the negatives and
create more complex phrases or sentences.
There are many studies that reveal second language
learners also pass through sequences of development like
first language learners, and the process that L2 learners
acquire their L2 is very similar to that found in L1
acquisition. One of the examples is that the features
acquired in the early stages by one learner will be also
learned in the early stages by others even though there may
be slight time differences. There are, however, noticeable
differences in developmental sequences for L1 and L2
acquisition.
The apparent difference between developmental
patterns in L1 and L2 is the stability of the sequences.
While L1 learners will consistently all have a similar
order, the order of L2 learners’ acquisition is not always
clear enough to propose a general picture of developmental
sequences.
The point which should be considered is that there
are a great number of influences from L1 occurring in L2
acquisition, and Developmental Sequences in L2 will be
changed according to the learners’ L1s. Therefore, for
- 3 -
instance, L2 learners of English whose native language
shows somehow completely different processes from those
whose L1 is German. From this point of view, it cannot
determine that there is a universal Developmental Sequence
which applies to all language learners whose L1 is
different.
Furthermore, there can be gaps among L2 learners,
who are acquiring the same L1, and with different
environments such as family, educational background,
school, personal experiences, and so on. In second language
acquisition, to some extent, it depends on the learners’
characteristics and how carefully the learners will pay
attention to the second language systems, and this result
will greatly affect the researcher’s ability to analyze
developmental patterns that appear in the process of
learning languages.
About 40 years ago most linguists believe that the
incompleteness of second language acquisition was simply
due to their first language. At that time the major method
used to study language acquisition in this field was
Contrastive Analysis (Lightbown and Spada)3. Its aim was to
clarify the differences and similarities between L1 and L2
and sought potential errors that learners may produce based
on L1 interference. The idea was that where major
differences were found (for example if an L1 was SOV and
the L2 was SVO, then there would need to be a lot of care
and attention paid to this difference. If however, both
- 4 -
languages shared articles, then the need for heavy
instruction was not needed. This was common practice until
the end of 1960s and was shown in the many drills and
repetition exercises common at the time.
During the late 1960’s and early 1970s, a new method,
error analysis, was developed to make more detailed
analyses of second language acquisition. The belief was
that errors that are produced by L2 learners are not always
due to the first language, but also other elements. This
method is based on observations from cross-sectional
studies, which studied subjects at different ages, and
longitudinal studies which examined the language
development of a small number of the same learners over
time. Error analysis focuses on errors produced as a result
of the process that learners try to discover the rules and
structures of their target language, rather than assuming
potential errors.
Out of this kind of analysis we came to see that L2
speakers repeatedly made the same kinds of mistakes and in
similar ways. Moreover, the kinds of mistakes they made
seemed to be occurring regularly. For example, L2 learners
at one level of ability consistently said I goed to the
bank or I buyed a book whereby they are using a past tense
formulation rule consistently but wrongly. This suggested
that these errors were in fact a separate system from the
L1 and the L2 because the words goed and buyed are not L1
or L2 items. This system came to be known as Interlanguage.
- 5 -
The concept of Interlanguage was provided by Larry
Selinker4, who defined Interlanguage as “the learner’s
developing second language knowledge” (Lightbown and
Spada5). When we consider second language acquisition,
Interlanguage is a most remarkable function because it has
both some characteristics of the learner’s first language
and some characteristics from the process of language
learning. However, it is true that there are difficulties
conducting targeted studies to identify developmental
sequences in second language acquisition.
Generally, when researchers try to research
developmental sequences in both L1 and L2, they focus on
analyzing errors from free production either in
conversations or essays for example. We shall now look at
the differences between errors and mistakes to provide some
clear background for the studies of developmental patterns.
1.3 Mistakes and Errors
It is necessary to discuss the differences between
mistakes and errors because it is quite hard to distinguish
them when they occur in free production. The basic
processes of how mistakes and errors happen are described
in the following figures.
Figure 1: The process of making mistakes
- 6 -
Each category is a function, which is necessary to take for
language learning. Input means the information that a
learner obtains by listening and reading. Learning process
works to analyze the input, and output is a result from the
learner’s trials using acquired knowledge. Output is
usually in the forms of writing and speaking. Monitor
system is the function, which focuses on producing correct
output, and this system greatly depends on individual
learners.
In Figure 1, a learner produces utterances (shown as
an arrow) through the learning and output stages, and then,
this utterance goes through the monitor process. When the
learner notices that he made a mistake, he will self-
correct and the utterance will go back to the output stage.
For example, if a learner says, “She go to the school,” and
he will find that he made a mistake. He corrects it by
himself, and produces the correct sentence, “She goes to
the school.”
Figure 2: The process of making of errors
- 7 -
Learning Output
Monitor
Input
Learning Output
Monitor
Input
Errors have very a similar process to mistakes. The
process of making errors can be explained with following
example.
A: Can I lend your dictionary?
B: You mean “can I borrow it?”
A: Oh, we say borrow not lend do we? Can I borrow
your dictionary?
This conversation describes that A did not realize she made
a mistake until B pointed it out. Therefore, external
correction was necessary in this process.
As the above suggests, the biggest difference between
mistakes and errors is the correction: self correction
occurs in the process of making mistakes through production
monitoring while external correction is done for errors. In
other words learners may know they have made a mistake and
can correct them but they do not know they have made an
error.
Second language Errors can be divided into four major
categories due to their characteristics:
i) Developmental errors
ii) L1 interference,
- 8 -
iii) Overgeneralizations
iv) Simplifications
Examples of error
Developmental errors
(1) I told my mother that it is more
interesting than study.
This error is considered a developmental error because
these types of error often happen typically to the learners
whose L1 does not have a “tense agreement” system. This
Japanese learner produced “is,” instead of saying “was,”
because people do not need to consider that system in their
native language, which is Japanese.
L1 interference
(2) My brother also enters the badminton club.
L1 interference error is literally caused by the learners’
first language. This example may be easily understandable
for learners whose L1 is Japanese, because people use the
same word hairu when try to say both going into and
join/take part in in Japanese. This word usage in L1
influenced production in L2.
Overgeneralization
(3) …she will be surprising and exciting….
This sentence shows an example of overgeneralization, which
L2 learners use certain grammatical features even though
- 9 -
the occasion is not required. In this case the words
surprising and exciting should use the past participles
surprised and excited, however, this learner used the
present progressive.
Simplification
(4) Do you have [some] money?
Simplification is an error where elements of a sentence are
left out. This example shows that a learner skipped an
element, some, or any when she produced this sentence.
Developmental errors reflect “the learner’s gradual
discovery of the second language system,” (Lightbown and
Spada)6 and they are very similar to the errors that
children learning their L1 often make. On the other hand,
L1 interference is directly influenced by the learners’
attempts to transfer some of their L1 characteristics into
L2 learning processes. Overgeneralization errors occur when
learners try to “use a rule in a context where it does not
belong,” and simplification means to extract essential
elements of a sentence (Lightbown and Spada)7.
Errors consist of two bases according to when they
appear: competence and performance. “Competence” refers to
linguistic knowledge or understanding, which exists in our
ability, and “performance” means the actual usage of
language in speaking, or writing.
- 10 -
Each leaner has their own “competence,” therefore,
one can acquire certain grammatical morphemes faster than
others. Naturally, there are learners who can follow their
classes where mainly they learn in the L2, and other
learners who tend to be behind the contents of the classes.
In such cases, what can happen when these L2 learners are
asked to produce something (performance) during the class,
and also after the class such as some kind of homework?
They may supply some similar production during the class,
however the later they produce an item, the more different
results researchers may get. This is because ..
For instance, if the learners learn the third person -
s in the classroom, during the class they will produce very
accurate and similar productions. However, when they do it
for homework or examinations after a week or so, some
learners will say, she go to school, or Does Tom eats
lunch?, while others may retain their stable production
like she goes to school, and Does Tom eat lunch? These
kinds of gaps between learners are often observed.
One who has enough and rich competence to acquire L2
will show stable developmental patterns while the other
whose competence is poor may supply random pattern results,
which is difficult to conclude that both subjects have
experienced similar developmental patterns.
1.4 Free production
- 11 -
Investigating free production is one of the most
reliable methods to look at developmental stages in second
language acquisition. The difficulty of this type of study
is, however, that there is a great amount of missing data
that researchers want to have, in both analyzing speaking
and written production.
Verbal oral production can provide reliable and
spontaneous data, which are purely self-produced.
Therefore, the distinction between mistakes and errors can
be made clear. Compared to this, there is another type of
production which is written production.
Written production, for example from essays, seems to
be easier to obtain, hence, more convenient. However, this
type also has several inevitable disadvantages if we look
carefully.
1.5 Conclusion
In Chapter Two a simple study will be conducted
highlighting some of the main developmental patterns in
second language acquisition. The study will help to reveal
the main theme of this thesis which is to investigate the
validity and reliability of conducting studies which aim to
generalize developmental patterns with data obtained from a
simple case study.
In Chapter Three the results of the study will be
analyzed and discussed to discover some of the main
problems in conducting this type of research. Chapter Three
- 12 -
will also identify some of the guidelines that future
researchers should pay attention to when they try to
conduct more valid studies aiming to examine developmental
orders.1 Hatch, E.M., p. 18.
2 Lightbown and Spada, p. 174
3 Lightbown and Spada, p.72-73
4 Selinker, L. 1972
5 Lightbown and Spada, p.74
6 Lightbown and Spada, p.173
7 Lightbown and Spada, p.75
- 13 -
Chapter Two
The Study
2.1 Introduction
In Chapter One the basics for understanding
developmental sequences and its research were discussed. In
Chapter Two we will conduct a case study based on the
points which were explained in Chapter One. Then in Chapter
Three we will analyze these results focusing on some major
problems that influence the investigation of developmental
orders in second language learning.
This study was conducted in order to identify several
problems that occur when conducting studies of
developmental patterns. Although there are many different
kinds of methods which investigate developmental sequences,
it was obvious that the results from those studies did not
provide us with the entire picture of patterns for L2
acquisition. To discover the common problems of conducting
this type of research, obligatory occasion analysis was
used.
Obligatory occasion analysis is one of the most
common methods L2 acquisition researchers use in this kind
- 14 -
of research. This will be explained later. The study
provides detailed results from two subjects, and there is
also a discussion of this simple study.
The aims of this study are to discover whether
a) developmental sequences can be found in the data
of two subjects, and
b) these data can be collected reliably
2.2 The study
In this section the study will be described.
2.2.1 Subjects
Two subjects volunteered to supply data for the study
of their developmental sequences. These are Subjects A and
B. The subjects were a college student and a High School
student in Okayama, and this was a convenience sample.
Subject A was 21-year-old senior at university
majoring in English. She has been learning English since
the age of 12 when the English education usually starts,
and this is her 10th year of learning English. Before she
began to learn English at school, however, she had been
strongly interested in English and influenced by music and
movies in English. Therefore, English was not a completely
new resource for her to learn.
Subject B was an 18 year-old-male, going to a High
School. Similar to A, Subject B also started learning
- 15 -
English at the age of 12. When he was 14, he began to go to
a private English cram school which was focusing on English
grammar.
Subject A and B have studied English mainly through
classes at schools, and this research is based on their
notebooks or papers taken during those classes.
2.2.2 Instrument and Procedure
This study was based on the resources from Subject A
and B throughout the term of their learning. These
resources were the written free production such as
compositions, essays, and papers. These were chosen in
order to see the developmental sequences as was explained
in Chapter One. It is very important to focus on the free
production because the function of Interlanguage will
clearly appear in free production more than the restricted
or controlled production such as repetition and pattern
practice drills.
The errors found in these resources were divided into
four categories according to the specific characteristics
of errors: third person -s and copula, tense, and countable
and uncountable. Each subject provided their notebooks and
papers, and the researcher looked through those resources,
checking the numbers of errors in each category and looking
for the obligatory contexts.
2.2.3 Obligatory Occasion Analysis
- 16 -
Obligatory occasion analysis was used as the primary
tool for investigating the data produced by the subjects.
It is one of the major methods used widely in the studies
of L2 acquisition. In short, obligatory occurrences are
occasions when a certain form must be produced and is the
only correct form that can be used. For example, in this
sentence “It is a book” only ‘book’ not ‘books’ can be used
as the sentence demands the singular form. If the sentence
‘It is a books’ were supplied and the obligatory form is
‘This is a book’, the supplied form is wrong. In other
words the obligatory occurrence was not supplied.
By looking at the percentage of correct suppliances
for obligatory occasions we can see how well a learner has
a acquired a particular form. If the learner always
supplies the obligatory form we can clearly say the learner
has acquired it. This is less certain when the percentage
is much lower, such as at say 50% or even 90%.
According to Ellis (1994)8 referring to Brown (1973),
there are three basic steps to collect these data. First,
“samples of naturally occurring learner language,” which
are freely produced, are supplied. Second, obligatory
occasions for each specific category of L2 are examined:
negation, questions, grammatical morphemes such as tense
and copula, and so on. Finally, “the percentage of accurate
use of the feature” is compared with obligatory occasion
numbers, considering “whether the feature in question has
been supplied in all the contexts in which it is required.”
- 17 -
We shall now look some more at examples of these
obligatory and supplied occurrences.
(Example 1) Obligatory occasion in tense
Obligatory: The box has never been opened since 1900 .
Supplied: The box is never opened since 1900 .
(Example 2) Obligatory occasion in countable and
uncountable
Obligatory:
My brother broke glass of our room. –Uncountable
I don’t like wearing glasses…. –Countable
Supplied:
Finally, and I got three glass [glass cups]. –
Countable
We can see that in Example 1 the obligatory tense is the
Present perfect, but the present simple passive is used.
Therefore in this example the subject will score 0/1. In
other words for one occasion when the correct form was
required the subject did not supply it. Similar
calculations can be provided for example 2. Example 2 shows
the obligatory occasion for countable and uncountable
nouns. The supplied noun glass should be plural since the
learner mentioned about glass cups, however she used glass
[cups] as a material glass, which always remains singular,
and supplied by mistake.
- 18 -
Based on the idea of obligatory occasions, the total
number of suppliances per feature were identified as
follows. For the sentences that included (or should have
included) third person -s and copula were counted in the
first category. Secondly past, future, and present and past
perfect tense obligatory and correct suppliances were
counted. Nouns, except pronouns, proper nouns, and numbers,
were collected as countable and uncountable, as mentioned
before. To prevent overestimating these results, each
feature, especially tense and third person -s and copula,
was counted once per clause. The method of analyzing these
results is shown in Examples 1 and 2.
Example 1
The treaty was founded on the presupposition that there is
nuclear relation between the US and the former Soviet Union.
Example 2
If it were not for the newspaper which tells us world ’ s
daily news, we would feel our daily lives inconvenient.
The first example shows that there are two clauses in the
sentence. The treaty was founded consists one past tense
feature, and one noun (countable), therefore, there are two
elements to count in part A. This suppliance is
grammatically correct, and therefore, the subject will
score 1/1 in both categories: tense and countable and
uncountable. Same as this calculation, the second one
- 19 -
provides three clauses with tense, third person -s, and
countable and uncountable correctly. According to this
result, the subject will get 1/1 for each feature in
Example 2. All data supplied by two subjects were
calculated followed by this method.
2.2.4 Results
According to the research, each feature of the
developmental sequences of subject A and B was summarized.
Tables 1 and 2 show the distributions of errors from each
category.
Table 1. Total number of suppliances that we analyzed by
feature for Subject A.
Type Total
3rd person& Copula. 122
Tense 104
Countable and
Uncountable
151
Total 377
Table 2. Frequencies and Percentages of Correct Suppliance
for Subject A
Junior High
3rd
Grade
High School
2nd
Grade
High School
3rd
Grade
University
2nd
Grade
University
4th
Grade
S.R. % S.R. % S.R. % S.R. % S.R. %
- 20 -
3rd person
and Copula11/33 66.7 17/27 63.0 13/17 76.5 18/22 81.8 19/23
82.
6
Tense 7/21 69.2 18/26 69.2 15/22 68.2 15/20 75.0 13/1586.
7
Countable
and
Uncountabl
e
13/34 62.5 17/29 41.4 16/31 51.6 22/40 55.0 12/1770.
6
Total 31/88 54/82 44/70 55/82 44/55
N.B. The S.R. column refers to the number of correct suppliances
for that number of obligatory occasions. E.g. the first item shows
that 6 of the 11 obligatory suppliances were correct.
Figure 1. The percentage of correct suppliance by
grammatical feature over the 5 years for subject A.
0
10
20
30
40
50
60
70
80
90
100
Junior High3rd Grade
High School2nd Grade
High School3rd Grade
University 2ndGrade
University 4thGrade
3rd Person &CopulaTenseCountable & Uncountable
Subject A provided her resources over 10 years, and
- 21 -
the total number of suppliance among three grammatical
features was 377 (Table 1). Table 2 shows the numbers and
the percentages of correct suppliances at the obligatory
occasions. For example, the first item shows that 6 of the
11 obligatory suppliances were correct, and its percentage
was 66.7%. Subject A produced a total of 31 correct usages
out of 88 obligatory occasions in Junior High 3rd grade.
Figure 1 (based on the data in Table 1) shows the
percentage of correct suppliance over 5 years. Countable
and Uncounrtable shows obvious up-and-down between Junior
High 3rd, High School 2nd, and High School 3rd years,
compared to other two features.
Table 3. Total number of suppliances that we analyzed by
feature for Subject B.
Type
3rd person& Copula.
Tense
Countable and Uncountable
Total
Total
137
107
112
356
Table 4: Frequencies and Percentages of Correct Suppliance
for Subject B
Junior High
1st Grade
Junior High
3rd Grade
High School High School
- 22 -
1st
Grade
3rd
Grade
3rd person&
Copula.
25/36 75.0 18/40 45.0 18/25 72.0 31/36 86.1
Tense 16/18 88.9 27/40 32.5 15/20 75.0 23/29 79.3
Countable
and
Uncountable
21/24 87.5 24/33 72.7 14/29 48.3 20/26 76.9
Total 62/78 69/113 47/74 74/91
N.B. The ‘suppliance’ column by school year refers to the number of
correct suppliances for that number of obligatory occasions.
Figure 2. The percentage of correct suppliance by
grammatical feature over the 4 years for subject B.
0
10
20
30
40
50
60
70
80
90
100
Junior High 1st Grade Junior High 1st Grade High School 1st Grade High School 3rd Grade
3rd Person & CopulaTenseCountable & Uncountable
Table 3 shows the total number of suppliances over 4
years for subject B. The total number of suppliances is
356, consisting of 137 from third person and copula, 107
from tense, and 112 from countable and uncountable
- 23 -
categories. Table 4 also shows the frequencies and the
percentages of subject B, and Figure 2 shows the graph for
each grammatical figure. Subject B supplied more complex
behavior and less stability compared to subject A.
Countable and uncountable similarly goes up-and-down over
time while the other two features also show great decreases
and increases, which could not be found in Figure 1.
2.3 Discussion
This simple study provided us with several important
findings. First of all, it clearly showed that the actual
patterns of acquisition for L2 could not be obtained. Even
thought some predictions about developmental sequences in
second language learning can be possible, such as tense and
3rd person “-s” and copula will be acquired easier and
faster than nouns, the results from this study should not
and cannot be generalized into a whole picture of
developmental sequences in second language acquisition.
This was expected as was discussed in Chapter One.
Mostly, the correct suppliances of two subjects have
increased over the learning terms (Figure 1 and 2). Subject
A seemed to have quite a stable increase while subject B
showed more up and down. There are several reasons we need
to consider these result carefully. Firstly, subject A had
more stable competence and performance than B. Secondly,
subject A might have had richer input to develop her
understanding of her L2 structure. Thirdly, the quality and
- 24 -
quantity of data that subject A and B provide were
unbalanced.
The most noticeable point in these tables is that
both subject A and B experienced a large decrease on
countable and uncountable nouns, which are very difficult
for Japanese native speakers to master. Subject A got the
worst percentages of correct use when she was in High
School second grade, and also B got only 48.3% of correct
use in High School first year.
There are two possible reasons for this, a) the
amount of nouns that subjects had to deal with increased
rapidly in High School education and b) the subjects
started to explore and discover the rules of nouns,
producing language more spontaneously, therefore the number
of errors increased. This is well known as the intermediate
plateau whereby learners re-organize their language systems
at this stage and seem to show apparent and temporary
decreases in ability.
Another point to look at is the high percentages of
correct usages in Junior High School. This may be because
of the educational system in Junior High School. First, in
Junior High School the grammatical contents of English are
comparatively low and easy at Junior High School, and also
teachers often monitor learners to make them learn
“correct” English. Secondly, at this early stage of
learning second language, school teachers do not require
learners to produce much language in the way of free
- 25 -
production. Many exercises done in this period of learning
are mainly grammar drills. Therefore, there are few data
that can be useful to observe patterns of development, and
correct suppliances are produced under limitation or
controlled from outside in line with the teacher’s wishes.
2.4 Summary
This chapter discussed a study examining two subjects
in order to describe developmental patterns in second
language acquisition. Even though the scale of the study
was not large and detailed enough, there were some points
that need to be discussed regarding the research methods.
In Chapter Three the problems found throughout the process
of this study will be discussed.
8 Ellis R, p. 74-75.
- 26 -
Chapter Three
Problems Doing Developmental Sequences Research
3.1 Introduction
It is clear that developmental sequences are
considered one of the more important processes through
which all language learners go through as mentioned in
Chapter One. Consequently, the previous chapter described a
study which was conducted to highlight developmental
sequences in second language acquisition and reveal
problems occurring in the process of conducting the
research. This chapter will focus on and investigate some
of the major problems which were found in the previous
study. These will be categorized into two types of
problems: methodological problems and analyzing problems.
Finally, a brief guideline for researching developmental
sequences in second language acquisition, specifically for
case studies, will be given.
3.2 Methodological Problems
As mentioned in the first chapter, there are two
types of study: cross-sectional studies and longitudinal
- 27 -
studies, and in both methods, sampling is the first problem
to overcome before we can proceed with these kinds of
studies. In case study research, which tends to focus on a
small number of subjects and often is longitudinal, usually
one to two subjects are studied like the study in Chapter
Two. This may seem to be sufficient, however there can be a
problem when the results from these subjects are completely
different as shown in the study. Thus to get generalizable
results, it is important to consider the subjects’
backgrounds when sampling is conducted in order to lead to
more reliable results concerning developmental sequences in
second language acquisition in case studies.
Similar suggestions apply to group study research
that deals with a larger number of subjects in a cross-
sectional study. In addition, the relationship between the
number of the subjects and the reliability of the results
should be clarified, and by doing this, sampling will be
more specific and targeted.
Another concern is how long subjects have been
learning the language. There seems to be a strong case to
suggest there should be a threshold amount of learning
before one can conduct experimental studies on
developmental patterns. Especially in longitudinal studies,
the data researchers can get directly depends on the length
of time the subjects’ have been learning.
The most noticeable point here is that subjects are
all unique and different, therefore the data which are
- 28 -
supposed to show their developmental order may not be valid
beyond that learner no matter how long researchers collect
their data. One subject may supply remarkable data over 10
years of learning a second language while the other seems
not to develop L2 structure as clearly. This is seen in
Chapter Two, Figures 1 and 2.
There are other difficulties concerning the length of
subjects’ learning L2 when collecting data. First, it takes
quite a long time to get enough data to discover certain
patterns that researchers would like to have. In the
previous study, data over more than six years of learning
English were provided, however there were a lot of missing
data. If researchers want enough data to identify
developmental patterns, they have to spend both a lot of
time and a lot of money to collect data. This can include
recording, data storage, and so on.
Secondly, it is important to collect a great amount
of data frequently. This can be done by changing the points
or the questions to ask in order to get enough data to
accurately see the acquisition processes in second language
acquisition. A small amount of data means that there is
almost no validity and reliability to identify the
processes at work. Even if a great deal of data are
collected, it is laborious to reexamine the data over and
over because if the data were collected a long time ago
even the subjects may not recognize what kind of
productions they are, or whether they were errors or
- 29 -
mistakes. Furthermore, there can be limitations or controls
coming from the context where the data is collected even
though the researcher can get enough data.
In Chapter One, the significance of collecting and
analyzing free production was discussed. Free production is
very reliable data when it is determined as purely produced
by subjects themselves because there is a great possibility
that researchers can misjudge tutored production such as
teacher-monitoring of essays as free production, especially
in longitudinal studies. Such teacher monitoring will of
course affect the data.
There is also a big disadvantage in research dealing
with language production in that it is almost impossible to
record a subject’s every spontaneous utterance so that
researchers may collect good language data and patterns. In
addition, recording all productions is both expensive and
time consuming. Moreover, it is quite difficult to
recognize whether they are mistakes or errors because
written production can be erased, and researchers may
misjudge mistakes as errors and vise versa, because of this
difficulty.
There is also a possibility that subjects do not
correct their errors intentionally because they are lazy or
they just forget. Subjects may also avoid using some
grammatical features, which seem to be more difficult and
complicated to use. This is just a small list of possible
errors that learners may make which may confound attempts
- 30 -
to analyze recorded productions.
As pointed out before, longitudinal case studies look
at the same individual subjects over a period of time. The
older the data become, the harder it is to determine that
the data were produced freely by the subjects.
Consequently, the reliability of the study itself will be
lower.
Considering these points it seems to be very
difficult to collect enough freely produced data, which
researchers want to have in a longitudinal case study.
Moreover, a case study which focuses on too few subjects to
find out general and clear definition for developmental
sequences in second language acquisition. Moreover, these
problems are magnified in group studies, which have to deal
with many more subjects and with much more time. This makes
it difficult to expect reliable data.
3.3 Analyzing Problems
As well as methodological problems, analyzing
problems can occur very frequently, and are also difficult
to resolve. The first point to take into account when
analyzing the data is about the subjects. As suggested in
the previous section, subjects are often chosen
conveniently, therefore, the differences in their
background information, which will largely influence a
study’s results, tend to be given insufficient attention.
It is important, however, to notice this and how these
- 31 -
differences emerge and to recognize what kind of problems
they will create.
First, each subjects’ educational and language
learning background has to be discussed referring to the
study. These include the contents of classes, time span,
the method of teaching, personal motivation, and amount of
exposure. Because the two subjects in the study were
Japanese students, the results may reflect the educational
system in Japan. In Japan, English education in the early
stages, such as Junior High and High School, mainly aims at
the acquisition of English grammar, and eventually, a large
proportion of class content is spent on grammatical
exercises like translation or patterned drills rather than
speaking or listening.
Both subjects A and B experienced “grammar-based”
classes, however A took several classes which were based on
free communication with native English speakers over two
years. B, on the other hand, practiced English grammar by
going to a private cram school throughout his Junior High
years. Even though two subjects were in the same
educational system, they and their developmental sequences
were influenced by different class content or extra
curricula activities which seemed to be small but continued
over certain periods of time.
The time span when the data were created is also
important to consider. We can assume that the sooner the
subjects produced the data after learning certain
- 32 -
grammatical morphemes, the more often and accurate they use
them in their production. To investigate developmental
sequences, it can be said that the freer the data are
produced, the more apparent the sequences will be.
Therefore, production created during or right after classes
may not be reliable data because there is a large
possibility that it is greatly influenced by the contents
of the classes. Researchers need the data, which are freely
created, and from this point, production performed after
classes, or a certain amount of time later, is more
preferable (Chapter One). If 80% of the data was created in
the classroom, for instance, while 20% was performed as
homework, the data cannot be completely relevant. In the
previous study, however, it is difficult to judge when
those data were created because they have the same massive
problems mentioned in methodological problems section.
Concerning the problem of time span when resources
are produced, the amount of exposure is another key point.
The amount of exposure means how often subjects experience
second language learning. Clearly the frequency of the L2
learning strongly affects their understanding of L2
structures and their acquisition. If researchers hope to
see the general picture of developmental sequences in
second language acquisition, they need to find subjects
with approximately the same amount of exposure, which seems
to be very difficult.
- 33 -
The two subjects in Chapter Two had different amounts
of exposure. The numbers of classes in school through
Junior High and High School were basically the same (three
classes per a week in Junior High, and four per a week in
High School), however, as mentioned in Chapter One, B went
to a private cram school twice a week while A did not have
special extra English lessons. Obviously, their length of
exposure is different, and the results from their data
should not be discussed without taking into account this
difference between the subjects.
Secondly, the subjects’ individual characters
including their interests toward language learning is
another point to look at even though it is very difficult
to measure the subjects’ interests exactly. Lightbown and
Spada9 also mentioned, second language acquisition was
greatly related to the learners’ characteristics. These
were classified into five features - motivation, aptitude,
personality, intelligence, and learner preferences.
However, in this regard there were several problems when
collecting the data. Firstly, there is a lack of reliable
methods measuring these five qualities, and secondly there
are problems of misinterpreting or overgeneralizing the
relationships between the five aspects as well as language
proficiency.
The subjects’ environment is one of the elements that
create those qualities, and concerning this point, the
variety of input that the subjects have obtained in a
- 34 -
particular environment affects second language learning.
Researchers need to know what kind of input subjects have
experienced or can get because observation and the
hypothesis we can make from these processes are based on
what they read and listen. To see developmental patterns in
second language learning emerge naturally, quite a lot and
rich input is necessary. General developmental stages that
would apply to many L2 learners will not appear if input is
too hard or poor, and again, it is not possible to know
what input the subjects have had before their data are
collected.
3.4 Guidelines
Throughout this chapter, the problems of this study,
both methodological and analyzing problems, were discussed.
In both categories it was obvious that time greatly and
directly affected research results and caused difficulty
that made it difficult to identify a clear image of
developmental patterns. As described throughout the
methodological problems section, time prevents researchers
from collecting valuable data as errors or mistakes
especially in longitudinal case studies. In addition, a
longitudinal study is not strong enough to provide a
general picture of developmental sequences since it deals
with quite a few subjects. One of the difficult processes
is to identify whether they are “errors,” which can be
useful information to see developmental sequences, or
- 35 -
“mistakes” (Chapter One).
One of the suggestions to reduce these indefinite
elements for longitudinal study is to conduct another type
of study. For example, Marilyn Adams suggested a different
kind of research as her study design (Hatch)10. This may be
a pseudo-longitudinal study which is designed so that
one may work simultaneously with children of different
language-experience groups, each having utterances
representing various levels of development: a composite
picture of different levels simultaneously. (p.279)
She combined a pseudo-longitudinal study and a longitudinal
study, and conducted a study, which was supposed to
discover more detailed and complete results in less time,
and better generalizations for developmental sequences.
In the case of Adams’ design, there were many
advantages that researchers could compare between different
L1 speakers and people who have different age. Moreover,
this type of research requires less, and more targeted,
time than usual methods such as a longitudinal study. There
are, however, some unresolved problems. Researchers may not
be able to see clear patterns or to get useful data. In
addition, there is still a difficulty with the subjects
themselves such as their background, their history of
input, personality, and so on. It is certain, however, that
such attempts to discover other study designs is very
- 36 -
worthy and significant in spite of the difficulties.
As described in Chapter One and Three, it is very
difficult to distinguish errors from mistakes, especially
if the time after collection is lengthy, and also,
collecting only errors is an incredibly time and cost
consuming process. Actually, the most efficient way to
collect errors based on free production needs some sort of
external pressure. For instance, if a teacher requires
subjects to write essays without using an eraser in classes
over certain periods, they may produce very useful data.
However, the learners may take more time to produce things
more carefully and this also can affect naturally produced
data. Even outside the classroom, a teacher or a researcher
can ask them to tape their own conversation for studies.
Without any external pressure or effort, it is almost
impossible to get variable data. However, subjects’
personalities and feelings must be taken account of the
results of their data.
3.5 Summary
Chapter Three analyzed some of the problems in
conducting a research dealing with developmental sequences
in second language acquisition. It also suggested
guidelines that future researchers have to consider when
they take this type of research.
It is very difficult to find a new efficient method
to provide a general picture of developmental sequences.
- 37 -
Since English education in Japan is based on the classroom,
researchers may find some difficulties to deal with
subjects. In other words, there is a great possibility to
obtain efficient and useful data if they can work with this
particular situation, specifically Japanese subjects.
9 Lightbown and Spada p. 76.
10 Hatch, p. 279-280.
- 38 -
Bibliography
Ellis, R. “Developmental patterns: order and sequence in
second language acquisition” The Study of Second
Language Acquisition . Oxford: Oxford University Press,
1994.
Fromkim, V. A. (Ed). Errors in Linguistic Performance—
Slips of the Tongue, Ear, Pen, and Hand-- . New York:
Academic Press, 1980.
Klein, W. Second Language Acquisition. Cambridge:
Cambridge University Press, 1986.
Lightbown, P.M., and Spada, N. How Languages are Learned.
Oxford: Oxford University Press, 1999.
Hatch, E.M. (Ed). Second Language Acquisition.
Massachusetts: Newbury House, 1978.