Before beginning our research with the student, we made several observations in his ESL classes at GSU, in order to gain some understanding of his proficiency level. The student is currently enrolled in a beginning level oral proficiency class, and intermediate level reading and grammar classes. We did have the opportunity to review some of his written work in the reading class. Our common feeling was that he had been placed correctly.
Our overall objective was to test the student's understanding of the English article system through a variety of oral and written production tasks. The objectives of our first meeting with the student were to find out background information, to have him articulate his overt understanding of the English article system, to determine methods with which he had been introduced to the English article system, and to determine his ability to distinguish between noun types from nouns in a list. The student who participated in this project is a 24year old Korean man who has lived in the United States for approximately four months. Before coming to the United States, he studied English in Korea for about nine years. He was able to contrast the differences in the two teaching styles that he has been exposed to when we asked him about the article system. He stated that in Korea, the article system had been presented to him as a memorization exercise. In the program at GSU, he said he had learned it in context. The student was able to articulate the main difference between definite and indefinite articles by stating that they distinguished between "what is a special thing and what is a general thing. When we directed the student to give an explanation of how the indefinite article changes to definite in second mention, he was unable to articulate an explanation. We then decided to test his knowledge of second mention directly through the exercises that would be presented to him in the following sessions. Also, before the session, we had predicted that the student had probably learned through memorization of lists so we began with a simple exercise that tested his discrete knowledge of nouns of all types. He was able to categorize singular, plural and noncount nouns with the exception of two irregular plural nouns.
From this initial session, we gained information about the student's way of learning, his educational background , and the extent to which he could handle the English article system at his particular level in his interlanguage. The information gleaned from these initial exercises helped us to develop assessment materials for subsequent meetings in order to properly address his current language abilities for maximum benefit.
In the second session, we wanted to focus on his ability to distinguish between article types in modified cloze, error identification, and error correction formats. The second target area was to assess the student's use of articles in past time narrative through spontaneous written production.
We also administered a modified cloze exercise using text from one of the textbooks used in our discourse analysis. One of the cloze exercises was a short story on Thomas Edison taken from Improving the Grammar of Written English (Byrd & Benson,1989, p.191). His performance on this exercise showed significant weakness with article usage, missing a little more than half of the prompts (52%). His biggest weakness proved to be in the discrimination between when to use the zero article versus the definite article. This initial exercise would later be called into question when further such exercises, as well as production exercises (both oral and written) displayed much greater accuracy with the same structures. We wonder if the unfamiliar content of the exercise (Thomas Edison) contributed to the student's initial lack of success. The student did comment that the exercise was difficult because he said he did not understand what the author was trying to say. It might be said that the lack of content schema (and context) negatively affected the student's output (Carrell and Carson,1983). We had anticipated this possibility, and therefore chose a passage from a history text which dealt with Korean history, hoping to see a difference in production outcomes. In order to avoid the possibility that the text itself, which was not designed to assess article usage, would provide ambiguous prompts, we administered the cloze instrument to a native English speaker who had no knowledge of either the text nor our purposes. The native English speaker was instructed to highlight any prompts in which his internal grammar would have accepted multiple responses. Those identified prompts were removed from consideration when evaluating production from our student. We found that the student did slightly better on the Korean history cloze exercise than the previous cloze exercise on Thomas Edison (39% missed vs. 52% missed) with the main difficulty being an inability to place the definite article or the zero article before nouns.
The student was then given an exercise which we used to ascertain how the student fared at identifying errors with noncount nouns. Out of five errors that should have been identified, the student recognized only two. He also recognized noncount nouns, such as money, as incorrect without an article. On a similar exercise taken from Problem/ Solution (Byrd and Benson, 1994), the student was asked to insert three missing articles into a text. The student inserted all three correctly. It should be noted that the student took a great deal of time to complete these short exercises, averaging fifteen minutes per exercise. The two most time consuming of these exercises were the past time frame writing exercises.
We began the third session planning to assess the student's ability to use articles in General Truth time frame. We had planned two writing prompts which we felt were likely to make positive use of the student's content schema from the information we had noted in our informal conversations with the student. One of the prompts was intended to assess his ability to produce definitions and relevant articles. The student was asked to write a paragraph explaining and describing his favorite Korean holiday; the student, however, seemed to bring his own agenda to the meeting, and he refused to write on these topics. So we compromised on the topic and the student offered us a rough draft of a paper he had been working on in his reading class. It seemed to match criteria and since he had written it by himself that day less than one hour before, we did not feel that reliability would be negatively impacted. We also tested the student's use of articles in General Truth time frame using a modified cloze exercise that was adapted from the psychology textbook in our discourse analysis. Again, as the text was not designed for assessing article use, we administered the instrument to a native English speaker as indicated above. The student had problems with the same type of errors as with the past time frame cloze questions. Out of twenty eight questions, the student missed eleven (39% missed). Nine of these errors were the result of the student misplacing use of the zero article with the definite error. Only two of these errors were a case of the student misplacing the definite and indefinite articles. It should further be noted that there were inconsistencies between the text and the output from the native speaker. These items were removed from consideration even though the student's response matched the text.
Our final session with the student was intended to assess the student's article use and noun discrimination in oral production. The exercises follow the same general pattern as the written exercises in that one conversation featured prompts designed to elicit general truth responses and the other to elicit past time frame responses. The prompts were again selected to utilize the student's content schema. Our goal was to take representative samples of the student's unmonitored vernacular speech. He agreed to talk about rock and roll music (general time frame), plans for an upcoming trip to New York, and vacation trips he had taken( past time frame). The format was a taped conversation in which we generated spontaneous prompts to keep the conversation as authentic as possible.
Our analysis of student writing and oral production uses the same parameters as the textbook discourse analysis indicated. Further attention is given to errors in student use of the relevant structures. Data was collected on the following issues:
As indicated in Appendix C, the student showed significant difficulties with article use and noun discrimination in past time frame writing, with specific difficulties in identifying and using indefinite articles, as well as plural nouns.
Use of the General Truth time frame in written production seemed to pose fewer difficulties for the student, the sole exception being, again, use of plural nouns. There was some disparity between the results of the written discourse and the oral discourse sections. While the student displayed continued difficulty with the past time frame, he appeared to have no difficulties with the use of the indefinite article in spoken discourse. General Truth spoken discourse results followed written discourse results, revealing few deficiencies. By equating the history text with the past time frame, and the psychology text with the General Truth time frame, it was possible to compare the nouns and articles from our discourse analysis with the written production of the student. There were significant deviations in noun choice in the past time frame , as well as minor deviations in the General Truth time frame. The value of such comparisons, however, may be the subject of some further discussion. We feel it important to issue this caveat: the writing and speech samples are far too small to develop specific hypotheses with regard to this student's language skills in noun discrimination and article use. However, even given the sparse data, we feel there are some specific pedagogical applications which might help this student.
Based on the analysis of student errors, we have noted the following general errors which the student seems to be making at this stage of interlanguage development (Note: Unless stated, cloze exercises are not included in this analysis):
The student appears to have the most problem with both past time and General Truth production (oral and written). The student displays an accuracy coefficient of 0.69 with the use of plural nouns. Major problems appear to occur with substitution of the singular for the plural form, as well as improper omission or inclusion of articles with plural nouns. Some examples include:
So from that time I could not make any girlfriend and I gradually kept at a distance from the friends. (past time production)
My opinion is that we should not lose our warm heart. (general truth production)
Singular nouns in past time production proved to be the next most problematic issue for the student, with an accuracy coefficient of 0.8 , followed by past time use of noncount nouns, with an accuracy coefficient of 0.83. The main problem here seems to be a combination of avoidance strategies, such as attempting to combine both singular and plural by using the determiner some:
Some professor may be fired because of a new professor called the computer. (general truth production)
Noncount nouns appear to be significant problems for the student as well, however, the accuracy coefficient is much higher than other areas .087. We feel that the problems are significant enough to report, due to the high incidence of errors the student made in earlier cloze exercises. The problems with the cloze exercises appear to be those where the student attempts to put an article on a non count noun:
… grooved cylinder covered with a tin foil. (past time cloze) (Byrd and Benson,1989, p.191)
No one needs to tell you, that the emotions add 0 to your life, that in times of the stress, they can disrupt your life or save it. (general time cloze) (Myers,1996, p. 335)
The student appears to have difficulties in distinguishing when a proper nouns takes a zero article and when it should take a definite article. This was found primarily in the cloze exercises:
… the electrical vote recorder that was intended for use by 0 United States Congress." (past time cloze) (Byrd and Benson,1989, p.190)
Such was a prestige of 0 Chinese, however, that the Korean alphabet was not adopted officially until 0 midtwentieth century. (past time cloze) (Upshur et al, 1991, p.265)
There seems to be a propensity to use the definite article where the indefinite article is required in written production. This type of error, although present in written production, appears to happen much more frequently in the cloze exercises:
I just took and got on the train and just left to somewhere. I didn't have the destination. (past time frame in written production).
0 certitude and 0 stubbornness helped Edison become the innovative genius (past time cloze)(Byrd and Benson, 1989, p.191).
Mizuno (1985) states that SLA research, in particular the Morpheme Studies, provides us a framework by which we can view the acquisition of the article system in English. In later work, Mizuno further develops the theory to include the mental processes involved in article choice (Mizuno, 1993). Although Mizuno(1985) states that SLA research, in particular the Morpheme Studies, Mizuno's work involved only Japanese learners of English, it might be inferred that his observations work in this case for two reasons: Korean is an altaic group language that has similarities with Japanese (Upshur et al, 1991) and, that Korean, like Japanese, has no system of articles. In general, Mizuno's theory holds that:
1) There is an order of acquisition of the article system, by type of article choice in relation to the meaning of the noun to which the article is attached.
2) Acquisition of these linguistic features can be analyzed via the types of errors a student makes.
3) The types of errors correspond to the level of acquisition of the article system.
4) Salient pedagogical implications can be drawn from an analysis of student errors since these errors correspond to developmental sequence (Mizuno,1993a).
If a student is not at a developmental stage where learning a particular linguistic feature is possible, then instruction of that feature might well be a waste of time for both the teacher and the student. Mizuno(1993a) identifies the following types of errors in article production:
1) Cooccurence the use of an article with another determiner.
2) Wordorder inversion of an article with an adjective
3) Underextension omission of a required article
4) Overextension using an article when the zero article is required
5) Substitution using "the" for "a(n)", and vice versa.
It is thought that all beginning students make these errors. Mizuno (1985) postulates that in the early phases of language acquisition, students try to seek out one to one correspondences between the native language and the target language, and semantic discrepancies appear as a matter of course. Intermediate learners begin to develop some rules to article use, and the errors begin to disappear in sequence, as new rules are acquired in the interlanguage. Our job in this regard is to analyze our student's errors, determine what developmental phase he is in with article acquisition, and design a possible teaching plan to meet his needs.
The suggestions made in this section are to be construed as highly subjective. They are based on our opinions, which are based upon the time spent with the student as well as our interpretation of the limited data.
Given these limitations, we feel that the student's errors place him at an intermediate level of acquisition according to Mizuno's work in which he has developed an analytic learning set where his errors tend to involve either underextension or overextension (Mizuno, 1985). We also feel that he is tending to shift toward overextension.
From this analysis which shows that the student has problems distinguishing between the proper use of the definite article and the zero article, we would begin by having the student look at article use in contexts in which we have versed the student. The student would be engaged in inductively developing rules for why some nouns take the zero article or the definite articles in the given context. While this procedure might not yield usable rules for a multiplicity of contexts, it would help the student understand that there are variations in use dependent upon context. Focus upon this issue might yield sensitivity to context when producing spoken or written language. Barring fossilization, it is hoped that enough practice would lead to greater accuracy and automaticity in noun discrimination and article selection.