Analysis of Error Sources in L 2 Written English by Indonesian Undergraduate Students

This study investigates second language learners’ production in writing, aiming to understand common problems the learners frequently have with English. The study focuses on the analysis of errors in 100 essays written by 50 Indonesian undergraduate students of English, which is aimed at identifying the commonly occurring errors, classifying them according to a linguistic category, and explaining the causes of the occurring errors. The study shows that a large number of the occurring errors resulted from intralingual sources, the proportion of which account for more than 60% of the total errors. The results conform to a number of earlier studies in this research field which concluded that the majority of errors L2 learners make stem from intralingual interference. Another possible explanation for the phenomena is that because the subjects participating in this study are university students majoring in English, they had at least mastered the basics of English when entering university. As a result, they are more likely to make fewer interlingual errors than students at elementary level who tend to rely more on their native language when dealing with the aspects of the target language which they have not yet mastered.


INTRODUCTION
The making of errors is basically part of a process one experiences when learning another language. However, errors made during the process of learning will gradually reduce and would be likely to subsequently disappear as he or she learns from his or her previous errors. He or she uses errors in order to gain "feedback from the environment, and with that feedback to make new attempts that successively approximate desired goals" (Brown, 2000, p. 126).
Researchers and teachers came to realize that a thorough analysis of learner's errors would help them understand "how language is learned or acquired and what strategies or procedures the learner is employing in the discovery of the language" (Corder, 1967, p. 25). Many researchers in the field of SLA have conducted studies on second language (L2) learners' language production in writing and speaking, aiming to study learners' L2 performance in order to understand their common problems in acquiring the L2. The main methodology and procedures employed in this research field include contrastive analysis, error analysis, and investigation of interlanguage. Despite their different stresses on different aspects of learners' performance, these three approaches all attempt to explore L2 learners' language in order to understand problems learners have with the second language.
Error analysis, which is the focus of this research, is one of the first methods employed to investigate learner language. It developed an important method for SLA research in the 1970s. Despite criticisms made against error analysis concerning its methodological procedures and limitations in scope, it is generally accepted that the study of error analysis has made and is still making a substantial contribution to the research field of L2 acquisition (Ellis, 2008). Consequently, this study intends to investigate the main research question: "What is behind the errors produced by the L2 learners in Indonesia? In other words, what appear to be the sources of such errors?"

Errors
In general, an error can be regarded as a deviation from the standard norm of the target language (Smith, 1994). Here, the standard norm refers to the standard written dialect. Ellis (1994) argues that errors committed by L2 learners can be in either comprehension or production. A comprehension error occurs when the learner misunderstands a sentence or an utterance. For example, s/he is unable to distinguish the sounds /ei/ and /e/ in sentences such as the following: 'pass me the paper' and 'pass me the pepper'. Errors of this kind, however, have been neglected since, despite our being able to test comprehension in general terms, "it is very difficult to assign the cause of failures of comprehension to inadequate knowledge of a particular syntactic feature of a misunderstood utterance" (Corder, 1974, p. 125). In this study, in accordance with the main concern of error analysis, the focus is on production errors.

The Significance of Errors
An error is not simply regarded as something to be avoided, but rather it is seen as a valuable part of the language learning process (Stern, 1983). With regard to this, Lewis (2002) puts forward the idea that the learner never making errors never learns anything. Consequently, the approach to language teaching is now shifting from how to avoid errors in how to talk for learning (McDonough, 1981). Put another way, errors are considered useful for language learning in the sense that learners would be likely to learn from them.
According to Corder (1981), errors are significant in three ways: for the teacher, researcher, and learner. First, on the basis of the systematic analysis, errors provide the teacher with information about "how far towards the goal the learner has progressed and, consequently, what remains for him to learn" (Corder, 1981, p. 11). For the researcher, errors provide evidence of "how language is learned or acquired, what strategies or procedures the learner is employing in his [or her] discovery of the language" (Corder, 1981, p. 11). Lastly, for learners, errors are indispensable given that they can be regarded as devices by which the learner discover the rule of the target language; they enable the learner to test his or her "hypotheses about the nature of the language he [or she] is learning" (Corder, 1981, p. 11). In addition, Hanzeli (1975, p. 431) argues that errors can be important for learners in that they make them "understand the total language learning processes" in which they are involved.

Contrastive Analysis
Contrastive analysis, which was popular in the 1950s and 1960s, is "a set of procedures for comparing and contrasting the linguistic systems of two languages in order to identify their structural similarities and dissimilarities" (Ellis, 1994, p. 698). The underlying assumption of contrastive analysis was that errors that learners made were considered to be largely a result of the negative transfer, or the so-called interference, from the learner's native language (Lightbown & Spada, 1999). In addition, it was believed that similarities between the two language systems, i.e. the first and second language, would enable learners to acquire the target language with ease, while differences between them would be sources of difficulty for learners in their acquisition of the target language (Selinker & Gass, 2008). Consequently, from the viewpoint of contrastive analysis, the differences between the two language systems are likely to result in more errors the learner tends to make when producing the target language.
However, the assumption underlying contrastive analysis came to be challenged for researchers, and teachers came to realize that in reality not all errors committed by an L2 learner were traceable to his or her native language alone (Allwright & Bailey, 1991). In addition, "sometimes the errors contrastive analysis predicted were not found in practice" (Allwright & Bailey, 1991, p. 83). Because of that, contrastive analysis fell into disfavour, and the researchers in the field started to look at a different approach to the analysis of L2 learner errors.

Error Analysis
Error analysis is not new since it had been a part of language teaching for a long time, for example, the earlier studies of learner errors conducted by French in 1949 and Lee in 1957 (Ellis, 1994). The traditional goal of error analysis was for pedagogical purposes; "errors provided information which could be used to sequence items for teaching or to devise remedial lessons" (Ellis, 1985, p. 51). However, error analysis as one of the important approaches to investigate the features of a learner language was established in the early 1970s, replacing contrastive analysis (Wray & Bloomer, 1998).
Even though both contrastive analysis and error analysis are related to the investigation and study of L2 learner's errors, the basic difference between these two approaches, as stated by Spolsky (1979), is that the former is "most concerned with language description" while the latter with "language acquisition and learning" (p. 252). In addition, contrastive analysis differed from error analysis in that it "generated predictions based on the comparison of the native and target language" (Allwright & Bailey, 1991, p. 83), while in the case of error analysis, it studied the errors learners actually made when producing the target language. Selinker and Gass (2008, p. 517) define error analysis as "a procedure for analyzing second language data that begins with errors learners make and then attempts to explain them". The analysis of errors involves investigating elements in the learner's interlanguage regarded as erroneous in relation to the standard norm of the target language. Dulay et al. (1982) note that there are basically two objectives of error analysis. First, error analysis is aimed at elucidating "what and how a learner learns when he [or she] studies a second language" (Dulay et al., 1982, p. 138). In other words, error analysis can provide us with "a picture of the linguistic development of a learner" (Corder, 1974, p. 125) and enable us to understand "how second language data is processed by a learner" (Lightbown & Spada, 1999, p. 74). The second goal is to "enable the learner to learn more efficiently by exploiting our knowledge of his [or her] dialect for pedagogical purposes" (Dulay et al., 1982, p. 138).
After the types of errors had been described, the next step which was the most important part of this study was to go through all the categories in order to explain the errors. This step was concerned with categorizing the errors based on the sources of the errors. There are generally three main categories of the sources of errors agreed by scholars (see e.g. Dulay et al., 1982;Dulay & Burt, 1974;James, 1998;Richards, 1974;Selinker & Gass, 2008), i.e. interlingual, intralingual, and ambiguous errors.
Interlingual errors can occur due to the transfer of rules or structures from the native language into the target language or the literal translation from mother tongue into the second language as a result of a defect in the learner's vocabulary (Corder, 1981;Sattayatham & Honsa, 2007). Intralingual errors are those that are common regardless the learner's L1 background and are caused by the interference of the language being learned, thereby independent of the learner's native language (Selinker & Gass, 2008). Errors of this kind reflect the learner's developing L2 competence, and they are "similar to those made by children learning the target language as their first language" (Dulay et al., 1982, p. 165). With regard to ambiguous errors, such errors can be traced from the learner's native language interference, and at the same time, they occur as a result of the interference of the target language (Dulay & Burt, 1974).

Participants
The subjects involved in this study were all Indonesian undergraduate students majoring in English Education at the Teacher Training and Education Faculty at Syiah Kuala University. There were 50 students participating in this research. In total, 100 written essays of two text types, i.e. narrative and argumentative texts, were collected from the participants.

Data Collection
One topic for each text type was provided: the topic for the narrative essay was "the unforgettable experience that happened to me", and that for the argumentative one was "studying the English language in an English-speaking country is the best but not the only way to learn the language". For each essay, the subjects were required to write a text of approximately 150-250 words in length. The time allocated for them to complete both writing tasks was about 100 minutes, which was expected to give them an adequate amount of time to finish both pieces of writing.

Data Processing
The processing of the data analysis in this research consists of four stages, i.e. the identification, description, explanation, and calculation of errors. After a corpus of learner language had been collected, the next step was to identify errors that learners made in their language production. All the errors that had been identified were later classified into certain categories which are used as a basis for subsequent explanation. One common way used to describe learner errors is based on linguistic categories. According to Chan (2004), the classification of learner's errors based on the linguistic category taxonomy would benefit both learners and teachers. This is where learners "may be more able to generalize the nature or causes of their errors" and teachers may be capable of offering "more effective and corrective feedback" (Chan, 2004, p. 67).
With regard to the calculation of errors, a statistical formula suggested by Sudijono (2006, p. 43) was used, as follows: The formula means that P refers to the percentage of occurrence of a particular error type, f is the frequency of occurrence of a particular error type and N is the total number of errors.

RESULTS AND DISCUSSION
In this section, I will report and analyze the sources of errors. The first part of the section presents an overview of the sources of errors based on the error analysis of the subjects' writings. In the second part, the number and frequency of interlingual, intralingual, and ambiguous errors for each of the main linguistic categories are discussed.

Frequency of Sources of Errors
It can be seen from Table 1 that the proportion of intralingual errors is about 60% of the total errors in the subjects' writings, which means that the great majority of errors made by the subjects result from intralingual sources. Errors of this kind might also occur in language production by native speakers of the target language at an early stage of language acquisition. With regard to interlingual sources, the frequency of errors that can be attributed to interlingual errors is relatively low, accounting for less than 15% of the total errors made by the students. The results conform to a number of earlier studies in this research field (e.g. Dulay et al., 1982;Dulay & Burt, 1974) which concluded that the majority of errors that occur in the language production by L2 learners are due to intralingual sources, not interlingual. In addition, based on the error analysis of the subjects' essays, nearly a quarter of the total errors produced by the students fall into the category of ambiguous errors, caused by both interlingual and intralingual interference. As is shown in Table 1, apart from the three main sources of errors, there is one more category included, i.e. unclear sources. Errors that belong to this category are those with unclear and/or unknown meanings, resulting in their being linguistically unanalysable. For this reason, the discussions of the error sources that follow exclude this category. Having discussed the number and percentages of the sources of errors, I will now examine the distribution of the number and frequency of the linguistic main categories in terms of their error sources.

Sources of Errors of the Linguistic Categories
As can be seen from Table 2, the great majority of errors in orthographical and lexical categories were caused by intralingual sources, accounting for 98.5% and 78.9%, respectively. On the other hand, errors resulting from interlingual interference occurred least frequently in the orthographical category, with the percentage being 1.5%. In the case of ambiguous errors, which are the combination of interlingual and intralingual, they were identified in the grammatical category. As is shown in Table 2, ambiguous errors were the major source of grammatical errors, the proportion of which account for more than 40% of the total grammatical errors, which is higher than that of the other two sources. After looking at the overall picture of the sources of errors, I will now move on to examine in greater detail the distribution of number and frequency of the sources of errors in each of the three main linguistic categories including the sub-categories. During the discussions, the examples of errors taken from the subjects' writings are also given. It can be seen from Table 3 that almost all errors made by the subjects occurred as a result of intralingual sources, accounting for 98.5% of the total orthographical errors. Only 10 errors (1.5%) were caused by L1 interference, and no ambiguous errors were found in the subjects' writings. As is shown in Table 3, except for spelling errors, all errors from the other sub-categories were due to intralingual sources. Interlingual errors, while being a source of some errors made by the subjects, seem to have a minor role since only less than 5% of the total spelling errors resulting from this source, and as can be seen from the table the majority of the errors (about 95%) occurred due to intralingual sources.
Indonesian, the subjects' native language, "is written in the Latin (Romanised) script" (Yong, 2001, p. 283), which is similar to the English language. In general there are a number of similarities between the two languages in terms of the rules of punctuation and capitalization, for example, those concerning the use of a full stop, comma, question mark, and exclamation mark as well as the use of a capital letter for a proper noun and to start a sentence. For this reason, errors made by the students in the two sub-categories can be categorised as intralingual errors since the sources of errors do not result from the interference of the learners' native language but rather from intralingual factors such as incomplete application of rules and ignorance of rule restrictions, which are also likely to be made by children learning English as the first language. Examples of common errors of these types due to intralingual sources are: 1) We always speak english [English] with native speaker.
2) The environment round you, [ -] will force you to speak English.
In example (1), the error occurred because the learner used lowercase for a proper noun. The learner's failure to apply the capitalization rule seems likely to be caused by the learner's lack of knowledge of the writing convention. In the case of the error in example (2), the subject applied a rule where it does not apply. In other words, this type of errors occurred as a result of the learner's ignorance of rule restrictions.
As is shown in Table 3, the students made many errors related to spelling, and the majority of these were caused by intralingual sources, whose proportion was more than 95% of the total spelling errors. Interlingual interference was only a minor source of errors, accounting for less than 5%. Below are examples of the errors made by the students: 3) We have to correct pronounciation.
[pronunciation] 4) When I was in hospital the doctor cleaned my arm and my knee that was hurt by alkohol. [alcohol] One common spelling error found in the subjects' corpus is the use of the word 'pronounciation' instead of 'pronunciation' as in example (3). The subjects' previous knowledge of the verb form of the word, which is pronounced, seems likely to lead to their making of this error. It can, therefore, be argued that intralingual overgeneralization seems to be the source of the error. Regarding example (4), the error that occurred was likely to be caused by L1 interference. This is because the word alkohol is a loanword from English which has had some modification from the original word, where the consonant /c/ in the word is replaced with /k/ in Indonesian. As a result, the subjects may be misled by the Indonesian spelling of the word when producing it in written language. According to Table 4, of the three sources, two of them, i.e. interlingual and intralingual, were the sources of lexical errors found in the students' writings, while no errors in this category were caused by ambiguous sources. Based on the error analysis of the subjects' essays, the majority of errors made by the students can be attributed to intralingual sources, the proportion of which accounting for almost fourfifths of the total errors. Another one-fifth of these errors resulted from interlingual sources of errors. In the case of the 'use of word from L1' sub-category, all errors of this type were due to interlingual interference. In the wrong word choice sub-category, interlingual and intralingual shared the responsibility of causing the subjects' lexical errors. However, intralingual factors appear to play a major role in this sub-category since, as is shown in Table 4, the percentage of intralingual errors was almost three times as high as that of interlingual. As for the word combination sub-category, errors the students produced were all caused by intralingual factors. Examples of lexical errors are: In example (5), the error occurred as a result of the student's use of a literal translation from L1. The word 'follow' means mengikuti in Indonesian, and mengikuti can be used in Indonesian to mean either 'follow' or 'take' depending on the context. So, it can be argued that the subject's L1 interference seems to be the possible source of this error because, as can be seen from the example, the learner tended to literally translate that word from Indonesian, resulting in his or her wrongly choosing the word in the target language for that context. Another commonly occurring lexical error also caused by the interlingual factor is the learners' use of the verb bring instead of take as is shown in example (6). In the subjects' native language, only one lexical item bawa is used to refer to these two words. Errors as in examples (7) and (8) were due to intralingual sources. In example (7), the subject perhaps could not appropriately distinguish between such lexical items as secondary and second, and as a result, he or she tended to use them interchangeably. The error as in example (8) falls into the 'word combination' sub-category, and the occurrence of this error might have been caused by the fact that the subject seemed unaware that the expression, in fact, is fixed in form. He or she might have thought that since the noun fact is countable, it requires an article. As is shown in Table 5, all three sources shared the responsibility of causing the students' grammatical errors. Unlike errors of the first two main categories discussed earlier, errors of this category occurred more due to ambiguous sources, the percentage of which being more than 40%. Intralingual errors, accounting for 38.5%, were the second major source of grammatical errors, while the number of errors attributed to interlingual sources was relatively low, less than half of that of ambiguous errors.
According to Table 5, interlingual, intralingual, and ambiguous errors that occurred in the noun category were relatively similar in their percentages, 36.5%, 29.5%, and 34% respectively. Examples of commonly occurring errors of the noun category are: 9) Some of us came into [the] river. 10) We can do the [ -] useful things like watch the film in English subtitle. 11) However, there are several way [ways] for non-English people to learn English in their home country.
The 'omission of article' errors are those commonly found in the subjects' writings as is shown in example (9). The occurrences of these errors were likely to be caused by interlingual sources. This is because articles do not exist in Indonesian (Yong, 2001), and as a result of this, the students have the tendency to leave out articles in English. In example (10), the intralingual factor seems to be the source of the error. The students who made this error were aware that English has the article system, but they were not really capable of using them appropriately due to their limited L2 knowledge of the L2 article system, and consequently, these students tended to overgeneralize the article usage, resulting in their making of these errors during their production of the target language. Another common error that students made is related to plurality as in example (11). The potential source of this error seems likely to be ambiguous. The L1 interference is that "Indonesian nouns are not inflected for number" (Yong, 2001, p. 283), and as a result of the non-existence of a plural marker, the Indonesian students tend to drop plural endings after number or plural quantifiers. However, the omission of a plural marker can also be attributed to the intralingual factor as a result of the L2 learners' strategy of simplifying the rule during their producing the target language; the learner might have thought that the use of the plural marker, in this case, the quantifier several, itself is adequate to signal plurality (Larsen-Freeman & Long, 1991).
As is shown in Table 5, verb errors, whose frequency of errors was the highest in the grammatical category, were mainly caused by ambiguous sources, accounting for more than half of the total verb errors. Based on the error analysis of the students' corpus, the majority of ambiguous errors occurred in the 'wrong tense' sub-category. The percentage of intralingual errors was also quite significant, accounting for slightly below 40%. Unlike ambiguous errors, errors of this kind were identified in the various sub-categories of verb errors. In the case of the interlingual sources of errors, they were less than 10% of the total errors, which indicates that the influence of the subjects' first language seems to have a minor role in errors related verbs. Example (12) shows that the student dropped the copula be whose presence is obligatory in that sentence. Based on the findings of the data, errors of this kind occurred frequently in the subjects' essays. Interlingual interference seems to be the source of these errors because, unlike English, Indonesian does not use the copula be for linking a subject with an adjective phrase and/or adverbial. Because of the lack of the copula be in the subjects' mother tongue, the students have the tendency to leave out the necessary copula be when producing such sentences. In example (13), the misuse of are as a marker of simple present tense occurred due to the intralingual source of false concepts hypothesized; the students failed to fully comprehend a distinction in the target language. Examples (14) and (15) are both related to tense problems. In Indonesian, verbs are not marked for tense, and as a result, Indonesian students are likely to have difficulty mastering the complexity of the English tense system. Example (14) shows the tense error that occurred might be caused by the subjects' overgeneralization of the present tense verb form. However, the absence of tense marking in the subjects' native language might also lead to their making of this error. Therefore, errors of this type seem likely to result from ambiguous sources. In example (15), the problem is that the student appears to have difficulty in comprehending the usage of different tenses; he or she might have hypothesized false concepts, resulting in his or her using the inappropriate tense.
As well as noun and verb errors, other common errors are those related to the preposition category. As can be seen from Based on the error analysis of the subjects' data, the prepositions 'in', 'at', and 'on' seem to be quite problematic for the students to handle successfully. At times they make use of wrong prepositions such as the problem shown in example (16). On the one hand, the complexity of usage of English prepositions seems likely to cause errors of this type. The subject who made this error may have been aware that the preposition 'on' is commonly used with a specific day, or part of a specific day, and dates. This previous knowledge of the target language might have been likely to result in the students' overgeneralization of the rule. On the other hand, the lack of prepositions in the students' native language (Yong, 2001) and/or differences between English and Indonesian prepositions in terms of the usage and meaning might also have led to their making of such errors. It could thus be argued that both interlingual and intralingual factors seem to have a role in causing the errors. The 'addition of preposition' problem as in example (17) might probably stem from ignorance of rule restrictions; the learners tend to apply the rules to contexts where they do not apply. In example (18), the students omitted the preposition, causing the sentence to be erroneous. The occurrence of this error might be due to the learners' being unaware that a certain preposition is required. So, the learner's lack of L2 knowledge seems to be the factor that caused the error.

CONCLUSION
The objective of this study is to investigate second language learners' production in writing, aiming to understand common problems the learners frequently have with English. The study focuses on the analysis of errors in the texts written by Indonesian undergraduate students of English in order to investigate what causes the errors in their L2 language production.
In relation to the main question addressed in this study, the results show that intralingual sources apparently has the most crucial impact on these subjects' L2 production. The proportion of errors which is caused by intralingual interference accounts for more than 60% of all the total errors made by the subjects. With regard to interlingual sources, the frequency of errors that can be attributed to interlingual errors is relatively low, accounting for less than 15%. The results conform to a number of earlier studies in this area which concluded that most errors the L2 learners make can be attributed to intralingual factors. Another possible explanation for the phenomena is that because the subjects participating in this study are university students majoring in English, they had at least mastered the basics of English when entering university. As a result, they are more likely to make fewer interlingual errors than students at elementary level who tend to rely more on their native language when dealing with the aspects of the target language which they have not yet mastered.