Research on Three Error Types of Chinese Esl Learners’ Acquisition of Passive Voice

This thesis aims to probe into the relationship between language transfer and the three error types of Chinese ESL learners’ acquisition of passive voice. On the basis of analysis of the result of a self-designed test, I come to a conclusion that under-passivization is the main characteristic of Chinese learner’s passive acquisition and negative transfer has a strong effect on Chinese learners’ acquisition of passive voice.

Language Transfer has been a central issue in second language acquisition and language teaching for at least a century. In the 1950s, it was often deemed the most important factor to consider in theories of second language learning as well as in approaches to second language teaching. In the 1960s, its importance waned as learners' errors were seen not as evidence of language transfer but rather of "the creative construction process." In recent years however, a more balanced perspective has emerged in which the role of transfer is acknowledged and in which transfer is seen to interact with a host of other factors in ways not yet fully understood. Selinker (1992) claims in his book Rediscovering Interlanguage that language transfer is the essence of interlanguage theory and highlights the role language transfer plays in constructing learner's l2 system.
Views about language transfer have undergone considerable change. In cases where the target language differed from the L1 this would result in interference or negative transfer. In cases where the patterns of the L1 and the target language were similar, positive transfer would occur. Thus, the L1 could both impede and facilitate the acquisition of the L2. In order to try to predict when interference would take place linguists carried out elaborate contrastive analyses of the native and target languages. Transfer often did not take place when there were differences between the target and native languages.
Transfer is one term of psychology. In traditional accounts of language transfer, the research focus was placed on the errors that learners produced. Errors occurred as a result of the negative transfer of mother tongue patterns into the learner's L2.
Negative transfer, also known as interference, is the use of a L1 pattern or rule which leads to an error or inappropriate form in the L2. Negative transfer can be found at all the levels of language structure. A substantial amount of empirical work in SLA (Second Language Acquisition) research has been devoted to establishing to what extent errors are the results of transfer (i.e. interference).

Three error types of passive voice
We will make an overall presentation of the three types of passive construction errors, (over-used, under-used and mixed-up) committed by Chinese ESL learners and we will come to probe into the typical patterns of the three types of passive errors under investigation.

Classification of Three Error Types
In linguistics, the class of linguistic units is called a type and examples or individual members of the class are called tokens (Richards et al., 2002:490). In the error-tagged version of Chinese Learner English Corpus (CLEC) , the error tag for voice is Vp7, which is the type of verb errors and all the related cases are tokens of Vp7, indicating errors in the use of voice: active for passive or vice versa. So in this thesis, we will adopt this item Vp7 to express errors in the use of voice. Moreover, we select some incorrect sentences for further analysis from the test paper to show certain results for those are completely picked up from the test paper by the subjects with great care and intention manually.
As we have discussed above, there are three types of errors in the usage of passive voice: over-used passive refers to the over-passivization; under-used passive refers to the under-passivization; and mixed-up passivization.
The error of overused passive indicates that the non-passivable verbs are passivized. In these cases passivization is overdone and extended to a great extent even happens to more unnecessary cases. By contrast, the error of underused passive refers to that passivization is insufficient. Passivization is underestimated in these cases where passivization is needed. Mixed-up passive refers to three cases including the wrong form past participle of verbs, wrong usage of passive auxiliary be and the misunderstanding of its passive meaning and even go against common sense or logics.
In this thesis, Vp7 is subdivided into three which are represented as Vp7, O, Vp7, U and Vp7, M. We use Vp7, O to stand for over-passivization, Vp7, U for under-passivization and Vp7, M for mixed-up passivization which has three subcategories as well: Vp7, M1 as the wrong form of verbs related with passive voice, Vp7, M2 as omission of auxiliary verb(be), Vp7, M3 as misunderstanding, insufficient or unilateral comprehension of its passive meaning.

Data Analysis of Three Error Types
In the translation part, we can see the distribution of three error types most clearly and easily. As mentioned above, the total number of passive voice occurrences in translation part we are going to check is 1560, (we have 104 subjects in the test), among which 677 mistakes concerning passive voice errors are found. Some of their occurrences are presented as the following.
2.*The failure is resulted in your carelessness.
3.*It is accepted that this matter will no be happened.
Case 2 Under-passivization 4.*Some people suggest that the meeting should be put off.
5.*I hope this kind of thing will not happen again.
6.*America holds the election in October. 7.*I pulled out a tooth this afternoon.

8.* Where is your education?
Case 3 Mixed-up passivization Case 3.1 wrong past participles 9.*The election is hold in October in America.
10.*Many problems should be discuss in meeting.
11.*He is not interesting in present discussing question.
12.*The failure is cause by your carelessness.
13.*The children were took care of by neighbors when she was ill.
Case 3.2 omission of auxiliary 14.*It is adviced that the meeting will delayed.
15.*It's a honor to admitted to study here.
16.*Where are you been educated?
17.*There things will not happen again was be expected.
18.*My brother is being to work in Xizang.
Case 3.3 incorrect or unilateral meaning 19.*By well cared, the baby is very healthy. 20.*This kind of thing was hoped not happened again.
21.*There are many problems will be discussed in the meeting.
22.*This kind of cloth is good to wash. 23.*This kind of cloth is hard to wash.
24.*This kind of cloth is patient to wash. 25.*This event was got his attention quickly.

Subject Selection
In order to facilitate the process, a longitudinal and empirical research was made on 104 students in China University of Petroleum who came from two different grades. All these 104 subjects are Chinese-speaking English learners. We will classify them into two groups, that is, group A and group B. Group A is composed of 49 2006-year-enrolled students, while Group B consists of 55 2007-year-enrolled students. They are supposed to represent the average sophomores of college non-English majors and the average freshmen in terms of English level respectively. The freshmen have spent half a year learning English while the sophomores spent one year and a half in learning after going to this college. Among them, there are 85 boys and 19 girls. All of them aged from 18 to 22 years old. Those students learned new Horizontal English Band 1 and Band 3 respectively in their intensive English class. Only one student began to learn English in his nursery time, while 20 students began to learn English in the elementary school and 83 students began their English learning in junior middle school. That means 76 percent of the subjects have learned English for over six years. Furthermore, 19 students lived in cities, 38 students lived in towns and 47 lived in countryside before they went to this university.

Approach
In second language acquisition research, tests are generally used to collect data about the subject's ability and knowledge of the second language in areas such as vocabulary, grammar, reading, writing, metalinguistic awareness and general proficiency. A test is a procedure used to collect data on subject's ability or knowledge of certain disciplines. (Seliger, Herbert W. and Shohamy, Elana 1989).
In this chapter, a passive structure proficiency test is made to probe into the students' acquisition of passive structure. A questionnaire was made out as well to interview the subjects to get data related with their background and other personal information. This thesis is intended to apply, from an empirical research, both the qualitative and quantitative analysis methods. Therefore, we would like to set a few questions, based on which to start out research.
For the listed questions, two ways are employed to assess the answers. One is to count the frequencies of each choice in respective questions. The frequencies will be shown in percentage. The other way is to take note of mistakes or errors that the subjects committed to the answers listed. For the questions that are elicited in the questionnaire and closely related to the research objectives, we will compare them with the results made by subjects to draw conclusions.
Among the two groups, a total number of 104 students are involved in this test and all of them are science majors.

Research Instrument
A quantitative and qualitative study will be used in this research in which empirical study is dominant and descriptive study serves as an important supplement. The quantitative study is based on a self-designed test. Full score of the test is 100. So the test will work as the data collecting instrument, which is necessary to be presented in great details.
It holds five parts: (1) Grammatical transformation from passive structure and active structure.
(2) Grammatical transformation from active structure to passive structure.
The qualitative research or questionnaire is printed on the top of the test paper, which includes the following items: sex, age, major (options among science or technology major or art major), initial time to learn English, scores of English subject in the entrance examination to college and living areas before going to college. The subjects are asked to fill in the blanks with authentic personal information and without the requirement of writing down their names, which relatively helps to obtain more objective answers. We plan to get detailed personal information about the subjects. From their answers, we can tell an outline of their prior living and studying setting to see if their prior background will be concerned with present score somehow, and get necessary statistics to further our empirical research.

Test and Questionnaire
To carry out this research effectively, a well-designed test paper is required. It is mainly made up of five parts.
Part 1 concerns the transformation of four passive sentences, which tests learner's acquisition of the basic structure transform from passive voice to active voice, which includes the following patterns, subject + predicator + that clause, subject + predicator + object + noun as the complement, subject + predicator +indirect clause + direct clause, subject + modal verb + predicator. Besides, it includes testing the correct transformation forms of passive verbs and tenses.
Part 2 concerns the transformation of four active sentences, which tests learner's acquisition of the basic structure transformation from active voice to passive voice, that is, subject + predicator + object + adjective as the complement, subject + predicator + object + noun as the complement, subject + predicator + object + infinitive phrase, subject + predicator + that clause. Besides, it includes testing the correct transformation forms of active verbs and tenses.
In Part 4, the subjects are required to translate 15 Chinese sentences into English, which aims to test whether the sentence translation is affected by L1 transfer in the acquisition of their native language (Chinese).
In Part 5, the subjects are required to translate another 15 English sentences into Chinese, which aims to test whether the sentence translation is affected by L1 transfer in their acquisition of native language (Chinese) as well.

Testing Procedure
The test paper was distributed to different subjects and collected in December 2007. All those subjects were asked to attend the test and required to finish within 50 minutes in the usual class time. The process of handing out and recollecting those papers was manipulated by their own teacher of those subjects in order to minimize any possible negative effects brought about by the researcher's intervention to make sure this a common and objective test.
During this process, the subjects were not allowed to consult any reference books or material or information resources to ensure them to go to great lengths.

Determining Norms and Criteria Procedure
After collecting and classifying the handed-in test papers, we have to deal with the data in great details. The process was made in two steps owing to the two different groups. Before the collected data can be subject to further analysis, both in quantitative and qualitative ways, they should be well prepared in the following steps.
(1) Setting norms of assessing the subject's acquisition by searching out the wrong spelling and errors of usage of verb phrases.
(2) Determining the criteria by which subjects' answer can be assessed.
(3) Determining the ways in which the answers to the question, especially the translation question, can be assessed.
As this test is made to test the acquisition of passive structure, we strictly follow the criteria. After the norms are set, the frequency of errors will be taken down for the five parts. As all the subjects have filled the same the test paper and questionnaire, the mere percentage of errors committed by subjects of different proficiency levels is sufficient enough to show the different acquisition status quo of passive structure. Therefore, the data will be presented in percentage of errors respectively.

Data Collection Procedure
First we mark the paper to sort out the wrong and the right, then count their frequencies and note them down. The collected data have to be well prepared before they can be used as subjects for further analysis, both quantitative and qualitative. It will involve several procedures in this process in great details: (1) Reading through all the test paper and questionnaire to set norms of assessing the subject's performance (2) Mark the wrong sentence and multiple choice errors (3) Determine the criteria by which translation part will be strictly scored.
(4) Determine the ways in which the answers to the questions can be assessed.
Before making a series of comparisons about the distribution of passive occurrences, we need first to identify and check all the passive construction exercises. As the first step, lines containing any form of passive mistakes were extracted to make a mark, calculate the numbers of passing numbers, divide by the total number of subjects and get various percentages. After this was done, an EXCELL program was used to pick out the faulty points. Completing the EXCELL form, we got all the data useful to make a comparison. All the work was done manually. Most testing questions were strictly picked up from Zhang Daozhen's A Practical English Grammar.
After the correction work, we select some incorrect sentences manually with great care and intention from the test paper for further analysis to get certain results and findings for those are picked up from the test paper done by the subjects. All the following English sentences marked with * are selected from this test material.
For the listed questions, two ways are employed to assess the answers. One is to count the frequencies of each choice in questions. The frequencies will be shown in percentage. The other way is to take note of mistakes or errors what the subjects committed to the answers listed. For the questions which are elicited in the questionnaire and closely related to the research objectives, we will connect them with the results made by subjects to make comparisons.
Among two groups, a total number of 104 students are involved in this test and all are science majors. Full score of the test is 100.

Data analysis
In this part, we will focus our attention mainly on the analysis of this empirical research. First, we will obtain the general descriptive statistics of the percentage of the passing number in the passive proficiency test for both groups. We can obtain the following general statistics on the variants of ESL learner's passive acquisition percentage for both groups of subjects.
The distribution of frequency counts of three error types will be shown in Table 1.
Insert Table 1 Here From the table above, we can have the major findings on the distribution of three error types: (1) Under-passivization and mixed-up passivization are two main passive errors committed by Chinese ESL learners.
(2) The most favored way for those subjects to express passive notion is under-passivization and it takes more than half proportion in all the Vp7 errors.
Under-passivization errors happen most frequently and are prominent in the three types of passive errors committed by Chinese ESL learners.
(3) The least frequently utilized passive construction is over-passivization. The errors of over-passivization just take a small proportion in the three types of passive errors.
(4) Mixed-up passivization errors rank the second and take a considerable great proportion in the three types, in which the wrong form of verb in passive sentence and the incomplete comprehension or misunderstanding of its passive meaning.
(5) Both groups have the similar results of the three error types, in which both take similar proportions.

Insert Table 2 Here
From the data in Table 2, we may get the following findings: (1) The average passing percentage of group A is 84 percent while group B is 71 percent, i.e., 47 freshmen passed the test while 35 sophomores passed the test. Compared with group B, group A has a higher percentage of passing rate. In other words, freshmen did better in the passive structure proficiency test.
(2) From the statistics above, we can see easily that 78 percent of the subject passed the test on the average, i.e., the majority had a good command of passive structure.
(3) They had a high passing percentage in Part 2 and Part 5 and 80 percent passed. In P4 (Chinese to English translation), they got the lowest percentage (43%), furthermore, both groups got the lower scores.
In terms of finding the general distribution, we make a vertical comparison or intra-group comparison. Let us take group A as an example. By comparing the frequency of the three types of passive errors within group A, we get the proportion of each type of passive errors to the total errors in this group. Thus the distribution of the three types of passive errors in group A is acquired. Adopting the same method, we get the distribution of the three types of passive errors respectively in Group B. All the statistics can be found in Table 5.2. On the basis of the above result, it is very easy to acquire the statistics on the general distributions of the three types of passive errors in the test.
In this thesis, major findings of present study will be summarized. Some limitations will be pointed out. Then some suggestions for further research will be presented. Based on the data analysis above, we come to the final conclusions.

Major Findings
We have discussed the subject's performance in the test and questionnaire. In the three error types, we find that under-passivization is the most frequent occurrence in three passive error types, next is the mixed-up passivization while over-passivization happen the least. What are the possible causes for this language phenomenon? We get the following major findings.
(1) Negative transfer mainly resulted from the difference of basic structure and system between Chinese and English language. As we have discussed, bei structure is the typical Chinese passive structure. Chinese has no inflectional change, which means Chinese verb would keep unchanged in different passive structures. English is the opposite. English verbs have corresponding inflectional changes according to specific grammatical requirement. As a result, this phenomenon may have a negative effect on Chinese ESL learners.
If the Chinese learners can not use these two languages freely, say, they can not find the corresponding expression structures in English. They may slide away from the option of English passive structure intentionally because they are afraid of making mistakes, which may frighten them into escaping from choosing passive voice. Also this is the reason why under-passivization is the most frequent occurrence of the three types.
(2) Subjects have a poor command of verb usages and structure. In certain or given patterns, Chinese learners are more likely to make mistakes because of their literal or word-for-word translation. They have difficulty in comparing past participle with present participle.
(3) Different features between English and Chinese passive voice make it hard for Chinese to distinguish from the two voices. Chinese and English belong to two completely different language systems. English verbs have different forms in active voice and passive voice, while the forms of Chinese verbs have not any change in different voices and tenses. During the acquisition of English passive, Chinese learners will inevitably commit errors because of the differences between the two languages (4) We know that there are great comparisons between the western and oriental cultural background so that the different thought patterns contrast. The Chinese care for the image, synthetic and parataxis thought, and the English is just the opposite, that is, they prefer abstraction, analysis and hypotaxis. Generally speaking the Chinese cultural influence and thinking patterns by and large cause a negative impact on English L2 learning.
(5) According to the Markedness theory, we can find certain relationship between Markedness theory and L1 transfer shown in Table 3.2, the tested past participle here works as a good case: L1 unmarked, L2 marked and interlanguage unmarked. Accordingly, the subjects are liable to transfer the unmarked patterns into their interlanguage.

Limitations of Present Study
During the process of research, we have drawn important conclusions on the acquisition of passive voice and have tried to be objective and scientific. However, there are still some limitations in this study.
We probe into the research only on two groups of subjects (104 people involved) and in the lay-out of research paper, although we have tried to cover all the passive types we could list as many as possible, it is likely that there may be some types left over or we did not cover.
In addition, in the questionnaire part, we only list six basic items about the subject's personal information. If more specific questions are made out, we may get more detailed analysis about the subjects. In the test part, if we could make out a more comprehensive testing paper, we would have had more detailed analysis and conclusion.
We conclude on the basis of a self-made test paper and questionnaire if we could collect corpus statistics from other aspects, the research scope may be enlarged to a greater extent.

Suggestions for Further Research
Since there are some limitations, there must be the necessity to do further research on Chinese ESL learner's acquisition of passive voice. Here are some suggestions.
We have only 104 subjects involved in this study. In the future research, the study can be put on a larger scale and more subjects could be selected consisting of more groups ranging from beginners, intermediate beginner and high level learners till the native-like level learners, to gain more objective results. A longitudinal study on more subjects is feasible and necessary.
Researching scope of the test could be expanded to a greater extent. Error types about passive structure could be subdivided into more details. Of course, many other fields concerning the acquisition of passive voice of Chinese ESL learners need further discussing and researching.