A Phonological Analysis of Errors Made by Iranian Intermediate EFL Learners: A Focus on Prosodic Features

Mehdi Mirzaei, Seyed Mahdi Araghi and Mohammad Reza Hasannejad, Iran

Mehdi Mirzaei was born in Tehran, Iran in1981. He got his BA in English Language and Literature from Isfahan University, Iran in 2004. He obtained his MA in TEFL from Iran University of Science and Technology in 2007. Currently, he is a Ph.D. candidate and is a faculty member in Islamic Azad University of Khodabandeh. His research areas of interest are bilingual education, the role of mother tongue in L2 teaching, and discourse analysis. E-mail: mmirzaei@khiau.ac.ir

Menu

Abstract
Introduction
Review of literature
Methodology
Participants
Instruments
Procedure
Results and discussion
Conclusion and implications
References

Abstract

The present study is an error analysis of prosodic features of Iranian intermediate EFL learners' speech. The analysis of errors and specifically the influence of mother tongue on the way learners shape the suprasegmental characteristics of their speech can help language teachers have a fuller awareness of the matter. In this study, the speech of 64 Iranian intermediate learners was analyzed in terms of the influence of Persian language on the stress and intonation of English words and sentences. The results revealed that part of the erroneous prosodic articulation produced by participants was due to the negative transfer of Persian phonological system. However, in those types of sentences which had a similar type of intonation to Persian, the errors proved to be the least. Consequently, learners’ resort to first language phonological system, in the absence of sufficient training in second language suprasegmental features, is one of the major reasons for first language-like rhythm of their speech.

Introduction

The present study aimed at investigating the prosodic errors Iranian EFL learners make while speaking English. Although most of the studies related to teaching the phonological aspect of second language focus on the accuracy of pronunciation, stress and intonation play a vital role in intelligibility of speech. These prosodic features are meaning distinctive; that is, an error in the correct use of stress or intonation may totally change the meaning of the sentence. Therefore, identifying the areas that learners make such errors can help teachers better understand to predict those areas and how to deal with this problem.

Speaking a foreign language is not a unidimensional skill; however, it consists of both segmental and suprasegmental components. The segmental component of speaking refers to the proper pronunciation of individual sounds including vowel and consonants. Nonetheless, suprasegmantal component deals with those features of articulation which are beyond the mere correct pronunciation of phonemes but those which consist of stress and intonation; that is, prosodic qualities of speech. In Skandera, & Burleigh's (2005) terms, suprasegmantal phonology, also called prosody, is concerned with those features of pronunciation that cannot be segmented because they extend over more than one segment, or sound. Such features include stress, rhythm, and intonation (also called pitch contour or pitch movement).

One of the major difficulties that Iranian English learners face in speaking is that their speech sounds more Persian than English. This is in spite of the fact that they pronounce most of the individual phonemes properly. Nevertheless, they fail to observe the rules of stress and intonation of standard English language which results in weird and Persian-like speech. It looks like replacing Persian words with English ones in the framework of Persian prosody. Consequently, the interference of first language seems to be the chief reason which accounts for this confusion. The present study tried to unveil the spots where and how first language resulted in these confusions.

In the field of language teaching and learning, error analysis can serve as a means to investigate the areas where language learners are most likely to make errors in comprehension or production of language. It can help teachers, to some extent, identify and/or predict why, how, and where their learners produce erroneous language. Therefore, errors made by learners in the process of language learning are not considered as a negative aspect of language learning but a natural step in development of language skills. Dulay, Burt and Krashen (1982) believe that analysis of errors made by language learners can help understand the process of language learning deeply; moreover, it will help teachers and curriculum designers to decide on teaching materials which best fit learning needs of language learners.

As a fundamental part of contrastive analysis, error analysis was used primarily to help language teachers predict what problems a language learner would have due to the linguistic differences between the learner’s native language and the target language. Errors that could not be attributed to language interference were virtually ignored and those most frequently focused upon, such as the omission of articles before unique nouns or phonological errors, were so well known that many teachers found the work of the researchers redundant (Ellis, 1985).

Errors are not confined to what is produced in language, but also comprehension of what is said or written. If we suppose that in the production of second language, part of the explanation for the occurrence of errors made by learners is due to the interference of first language, then we could regard the errors in comprehension as to be related to the schemata which have been shaped through the learners' first language experience. However, in this study, we analyzed the phonological errors Iranian learners made in speaking, in general, and prosodic features in particular.

Review of literature

Cook (1992) argues that any attempt done by the teacher, L2 learners already possess their mother tongue knowledge which is at times connected with their L2 knowledge. Consequently, it is not rational to regard L2 learning separate from learners' L1. Moreover, Selinker (1992) stated that there is negative transfer of L1 to L2 structures which naturally leads to errors especially when there are structural differences between L1 and L2; however, the transfer is positive in case of similarities between L1 and L2 structures.

According to Lado's theory of Contrastive Analysis, transfer was regarded as the most significant process occurred in the process of second language learning because of the relation it has with learners' first language. He argued that acquiring the similar elements between L1 and L2 structures is easier whereas the acquisition of different elements prove to be more problematic or even impossible.

Markedness is another crucial factor that must be taken into account when talking about substitutions. Since CA was not reliable to predict the order of acquisition or the level of difficulty, something else should be involved. Eckman (1977) proposed the Markedness Differential Hypothesis (MDH) and claimed that this theory could predict facts that CA could not. According to him, the most frequent features in languages are less marked and easier to acquire, and the least frequent ones are more difficult and; thus, more marked. Because the MDH was based on universals and on the relationship between L1 and L2, it was predicted that the L2 learner would have more difficulty to acquire a more marked form that is not present in his/her L1.

Transfer of linguistic items, including word order, certain vocabulary and sounds, from first language to second language is another process involved in second language acquisition. Selinker, quoted in Ellis (1994) reports that transfer may also be "responsible for fossilization" (Ellis 1994, p. 309), meaning that certain L2 sounds are consistently replaced by a phonetically close LI sound; the correct L2 sound is not incorporated into the interlanguage. Transfer is an important source of errors in second language, but not the only one, although it is more obvious at the level of the sound system than at other levels of language. For example, some phonological errors that were assumed to be the result of LI interference in later research in child language acquisition and second language acquisition were shown to be developmental errors.

Skandera, & Burleigh (2005) believe that intonation has some basic functions such as: a) the structural function signals the grammatical or structural role of an utterance and establishes its function such as interrogation, request or instruction. b) accent affects the emphasis of a syllable and therefore affects the focusing stress on particular words in connected speech. c) attitude is also transferred through intonation. It conveys speakers' personal point of reference towards their utterance and signals speakers' feelings. d) turn taking process is another function of intonation which occurs in exchanges.

Beh-Afarin, et al (2009) studied the effect of oral dialogue journals on Iranian EFL learners’ pronunciation. The participants were randomly assigned to treatment, control, and placebo groups. In the treatment group, the participants responded to the researcher's questions for two to five minutes and their speech was recorded. On the other hand, the placebo group recorded monologues and the control group's pronunciation practice was according to their course book. The participants' performance was corrected by the researcher. It was concluded that those practicing the treatment outperformed the control group in pronunciation in the areas of word stress, connected speech and intonation.

Sorayaie Azar and Molavi (2013) relying on the findings of their study on 13 EFL learners, tried to investigate EFL learners’ attitude and preferences toward different types and methods of error correction. Their findings suggest that EFL Learners are strongly positive toward teacher error correction of any type. They also indicated that Grammar and Phonology errors were the mostly preferred to be corrected and self-correction was the mostly preferred method.

Investigating the effects of L1 on L2 learning, A. Ridha (2012) tried to analyze EFL learners’ major difficulties in writing considering the nature and distribution of writing errors. He also attempted to find if there is a link between learners’ L1 and their English writing skill. In his study of 80 EFL university students, he collected learners’ writing samples and classified their errors into (1) grammatical, (2) Lexical/ semantic, (3) mechanics and (4) word order type errors. The results indicated that the majority of the learners rely heavily on their L1 while representing themselves through English. It also showed that grammatical and mechanical errors are the most frequent and sever types of writing errors among EFL learners.

One of the common difficulties in teaching pronunciation is the limitation of invested time and also the testing problem. Pennington (1998) believes that the majority of phonetic instructions do not positively affect learners’ pronunciation, since most of them generally focus on short periods of sounds instruction. He noted that in case of testing pronunciation, rater should be one who (1) is professional in phonology, (2) well-informed about testing criteria, and (3) utilize testing procedures similar to those applied in training. While advocates of naturalistic methods (e.g, Comprehension Approach) think that listening should be taught before speaking, it is believed that pronunciation be considered as a vital component of language learning pedagogy (Leather, cited in Chela-Flores, 2001). Levis (2005) put it in this way: “pronunciation pedagogy has always been investigated by ideology and intuition rather than research” (p. 369).

Responding completely to all the remarks rather than giving short answers, Ho (2003) showed concern about the features of teachers’ reply on oral journals. In addition, when learners request guidance, the instructor has to reply and to be helpful and sensitive to learners’ answers. Station (cited in Ho, 2003) pointed out that learners should think that they are involved in an informal discourse rather than participating in a formal testing task. As the emphasis is on communication, learners’ errors should be corrected based on intelligibility or when they are iterated repeatedly.

Methodology

Participants

In order to achieve the objectives of this study, 80 participants studying at Islamic Azad University of Khodabandeh and Arman Language Institute of Khodabandeh took part in the research. The participants were intermediate learners of English and their age ranged from 12 to 24 years old. Also, the participants consisted of both male and female learners. They were given a homogeneity speaking test in order to select a group of learners who were almost at the same level of speaking proficiency--intermediate level. Having collected data from this test, 64 homogenous participants were selected for the study.

Instruments

In order to carry out the study, a teacher-made speaking test was developed in the format of interview. In fact, the reason for developing such a test was twofold; in the first place, it functioned as the homogeneity test through which were able to select participants of almost the same level of speaking ability, and secondly, it provided the data we needed to analyze the errors they made regarding the prosodic features of their speech. Consequently, the only instrument used in this study was a homogeneity speaking test that provided the researchers with plenty of sentences to investigate their accuracy of stress and intonation usage according to English phonology system.

Procedure

The procedure for conducting this study had three main parts. The first part was designing a speaking test which consisted of questions which ranged from elementary to advanced-level of proficiency to be answered. The questions were designed by the researchers and consisted of questions which were related to the personal life of participants or those of everyday issues with which they were familiar with. In addition, the researchers wanted the participants to ask a few questions in order to have statement, negative and interrogative types of sentences for the task of error analysis.

The second stage was to conduct the test and rate the participants according to their performance on the speaking test so the researchers could select them as intermediate-level participants. To do so, all the 80 participants were interviewed individually by two raters and their speech was recorded for the purpose of rating and analysis.

Having interviewed the participants, FSI scale for rating speaking ability was used to rate the performance of the participants by two raters. In order to achieve inter-rater reliability, a correlation between the scores given by the two raters was conducted and a reliability of 84% was achieved. Based on the gained scores, the speech of 64 homogenous participants who were identified to be intermediate learners was used for the purpose of the analysis in the study.

The main task of analysis of the speech samples had just started. The researchers carefully listened to the recorded data and examined them only in terms of prosodic features. In other words, the pronunciation of individual words was not considered in the phonological analysis in this study, and only the prosodic features; that is, stress and intonation of the sentences were investigated. In addition, other aspects of proper sentence formation such as grammatical, discoursal and semantic characteristics were not considered in this analysis. Therefore, the error analysis of the speech samples had two major subcategories, stress and intonation. As for stress, the one, two and three and more-syllable words were identified and analyzed separately. The intonation pattern of sentences was divided into four types of sentences namely statement, negative, Yes/No questions and Wh questions.

To analyze the errors of stress pattern of each single word uttered by the participants, all the words except for prepositions and pronouns were taken into account. The total number of nouns, verbs, adjective, adverbs the participants uttered was 2240. Then all the words were divided into one-syllable, two-syllable, and three and more-syllable words. The descriptive table of the word types is shown in the table below.

Table 1. Word- type Stress Patterns

As mentioned before, the interviewers attempted to elicit as many types of sentences as possible. In fact, they sometimes wanted interviewees to ask them some questions as well in order to have different types of data for analysis. The total number of the sentences said by participants was 1088. The descriptive table of the sentence types is shown in the table below.

Table 2. Sentences- type Stress Patterns

Results and discussion

As pointed out in the procedure section, an interview was used to gather data for the purpose of error analysis on the phonological, in general, and prosodic features, in particular, of the Iranian intermediate EFL learners' speech. The results of the analysis revealed the following facts about the errors the participants made in stress and intonation of their sentences. The tables below describe the amount of deviation the speakers had in their speech from standard English prosodic patterns. The descriptions of stress and intonation errors are stated separately in the tables 3 and 4.

Table 3. Description of the Stress Pattern Errors

According to the table, the most erroneous area for participants was the stress of three and more-syllable words with the percentage of 61%. The second most erroneous type of words was two-syllable words, and one-syllable words were the words which were uttered with the least errors with the percentage of 21% and 0.5% respectively. Although a number of words were not said correctly in terms of stress, the pronunciation of them was quite understandable and the faulty utterance of the single words did not lead to major misunderstanding.

On the other hand, the error analysis of the intonation patterns of the sentences revealed that all types of sentences, that is statements, negatives, Yes/No questions and Wh questions underwent almost the same number of errors. However, the point that was common for the errors made in all types of sentences was that the participants did not seem to have been trained for the different intonation patterns of English language. For example, the fact that the words that carry the most of the sentence meaning receive more prominence was not something that could be recognized clearly in the speech of participants. Table 4 manifests the frequency of the occurred errors of intonation for the four types of sentences.

Table 4. Description of the Intonation Pattern Errors

percentage of 17%. Interestingly, the intonation of Yes/No questions in English and Persian are almost the same, i.e. in Persian, Yes/No questions have rising intonation. This similarity can be the main reason for the low rate of errors in this type of sentences.

Another interesting point which was noticed while analyzing both stress and intonation was the interference of first language in the stress and intonation pattern of the English words and sentences. Some of the errors made regarding stress showed that the mis-utterance of them was due to the subconscious replacement of the stress pattern of the Persian equivalent with the English words. In other words, some of the participants tended to say words with the stress pattern of the meaning of those words in Persian. For instance, the words "fifty", "seventy" and "eighty" have their stress on the first syllable, while their equivalents in Persian /pændʒah/, /hæftad/, and /hæʃtad/ respectively, are stressed on the second syllable. It was seen that some of participants placed the stress of these words on the second syllable the way they say them in Persian.

Moreover, cognates proved to be another source of stress errors. There are a plenty of words that have been imported from English to Persian. However, they have undergone changes in their stress patterns and have gained new stress according to Persian phonology system. For example, the word "internet" has a first-syllable stress in English; however, in Persian we pronounce this word with the stress on the second syllable. Therefore, it was observed that in 78% of the cases that this word was used by the participants, they pronounced it with the second-syllable stress.

The same interference of first language happened in the intonation of the sentences much more evidently. In most of the sentences said by the participants, it was easily recognizable that the Persian intonation pattern governed much of their speech as if they were speaking Persian with English words, or replacing Persian words with English ones in the framework of Persian prosody.

In addition, monotonous speech proved to be a major drawback in participants' speaking samples. In effect, in the speech of those with most errors and unintelligiblity, the lack of rhythm in their sentences seemed like saying each word separately without contribution to the flow of natural speech. However, English is a language in which tone has a crucial contribution to the meaning of utterance and the rise and fall of pitch can be meaning distinctive. Another shortcoming of participants' erroneous speech was that they did not notice the unstressed words in normal English sentences. Certain words in English such as prepositions, pronouns and conjunctions have a low pitch in normal sentences; however, in most of participants' erroneous speech, they received equal stress as other words such as nouns and verbs. Consequently, one of the prominent reasons for making monotonous sentences is stressing all the words equally and failing to recognize which words or phrases should normally be unstressed.

Conclusion and implications

The present study attempted to analyze the prosodic errors intermediate Iranian EFL learners made while speaking. In the discussion section, a comprehensive analysis of the errors regarding stress and intonation was presented. From the above discussions, two major points can be made regarding the explanation for the occurrence of the prosodic errors.

In the first place, first language proved to play an important role in leading foreign language learners to overgeneralize the prosody system of their mother tongue to the second language. In this study, which comprised the investigation of stress and intonation, it became evident that the participants' first language, i.e. Persian influenced the way the participants shaped their speech prosody. Both stress and intonation patterns seemed to follow Persian phonology system to a certain degree which made the speech sound more like Persian than English.

In the second place, the significance of proper training of suprasegmental features to the learners seems apparent. From what is attained from the mentioned findings, it is evident that L1 negative transfer is one of the sources that leads to erroneous speech, in general, and prosodic, in particular. This transfer seems to occur subconsciously until they haven't had received sufficient training in second language prosodic system. In other words, since the only resource L2 learners have at hand is their first language phonological system, it is not surprising to observe that they resort to that only resource as a foundation for building the phonemes, words, phrases and sentences rhythm of the second language on. Therefore, we cannot blame L2 learners for faulty segmental and suprasegmental speech until we, as language teachers, have provided them with sufficient training on those features. What follows is a few guidelines for language teachers to better prepare their learners for acquiring the suprasegmental features of the second language.

Firstly, language teachers can utilize listening to native speaker records as an authentic material for teaching phonology. It should be mentioned that most learners and even teachers approach listening as mere comprehension of the discourse; however, it can function as the main source for acquiring the segmental and suprasegmental phonological characteristics of language too, i.e. pronunciation of single phones and combinations of phones, stress pattern of words and phrases, and intonation pattern of sentences. In short, practicing listening should not be limited to the comprehension of the listening material, but attentions have to also be focused on the phonological aspects of authentic speech.

The second practice which can be helpful is repetition. Repeating after the recording makes learners identify the contour of stretch of speech and say the sentences with the intonation similar to that of a native speaker's. Of course, the teacher should center the attention of learners on how native speakers shape the contour of sentences and consequently, the imitation task will be more fruitful.

In the third place, reading aloud from a written text is another means through which learners can practice saying different types of sentences with correct stress and intonation. Again, in this activity of reading, comprehension is not the only purpose for reading, but articulating the sounds and sentences aloud in order to practice approximating the rhythm of words and sentences to that of correct and native-like one. Most of the teachers ask learners to read some parts of the course books aloud for the class for the purpose of checking their pronunciation of single words; nevertheless, focusing learners' attention on prosodic features of reading aloud the sentences will help them practice; and hence, internalize correct shifting of pith and loudness of voice in certain points of the sentence. Teacher's supervision and guidance is crucial in assisting learners to recognize and accurately articulate the suprasegmental components of sentences.

Finally, the findings of this contrastive research study can help both teachers and learners gain an awareness of the possible interference of mother tongue in the production of second or foreign language regarding prosodic features. This awareness will specifically make teachers sensitive to the fact that most learners will resort to their first language phonological patterns in order to frame the second language words and phrases. Therefore, teachers are advised to train second language learners about the specific phonological features of the language they are learning and differences between their first and second language should be clearly identified in order to avoid possible interferences.

References

Beh-Afarin, R., & Moradkhan, D. (2009). Oral Dialogue Journals and Iranian EFL Learners’ Pronunciation. Journal of Teaching English as a Foreign Language and Literature. 1, 21-37

Chela-Flores, M. (2001). Pronunciation and language learning: An interactive approach. International Review of Applied Linguistics, 39, 85-101.

Cook, V. (1992). Evidence for multi-competence. Language learning, 42, 557-591.

Corder, S.P. (1981). Error Analysis and Interlanguage. Oxford: Oxford University Press.

Eckman, F. R. (1977). Markedness and the contrastive analysis hypothesis. Language Learning, 27, 315-330.

Ellis, R. 1994. The study of second language acquisition. Oxford: Oxford University Press.

Ellis, R. (1985). Understanding Second Language Acquisition. Oxford: Progress Press.

Ho, Y. K. (2003). Audio-taped dialogue journals: An alternative form of speaking practice. ELT Journal, 57, 269-324.

Lado, R. (1957). Linguistics across cultures. Ann Arbor: Michigan University Press.

Levis, J. M. (2005). Changing contexts and shifting paradigms in pronunciation teaching. TESOL Quarterly, 39, 369-377.

Ridha, N. (2012). The Effect of EFL Learners' Mother Tongue on their Writings in English: An Error Analysis Study. Journal of the College of Arts. 60, 22.

Selinker, L. (1992). Rediscovering Interlanguage. London: Longman.

Skandera, P., & Burleigh, P., (2005). A Manual of English Phonetics and Phonology.Tübingen: Narr.

Sorayaie Azar, A., & Molavi, S. (2013). Iranian EFL Learners’ Attitudes toward Correction of Oral Errors. The European Journal of Social & Behavioral Sciences.

Pennington, M. C. (1998). The teachability of phonology in adulthood: A re-examination. IRAL. Quarterly, 11, 311-328.

Please check the Methodology and Language for Secondary Teachers course at Pilgrims website.
Please check the How to be a Teacher Trainer course at Pilgrims website.