Part I : Data Analysis

What is the overall extent of errors on verbal inflection in past tense in the data?

According to the tree structure, verbal inflection would include many tenses, i.e. simple past tense, past continuous tense and present perfect tense. The overall extent of these kinds of errors is at a very great level. Nearly all usage of verbal inflection in past tense is incorrect. To be specific, the error rate is staggeringly high (91.18% based on 31 errors out of 34 usages). The only three exceptions are ‘got’ in ‘...I got one’ (line 16), ‘was’ in ‘I go Chinese church when I was small’ (line 49) and ‘wen’’ in ‘We wen’ to a…’ (line 65).  All the other uses of verbs in past tense are wrong, including past simple tense such as ‘Then I transfer myself to Tomlin’ (line 11), present perfect tense such as ‘you ever been to China’ (line 28), and past continuous tense such as ‘When I am schooling in Primary Two to Six’ (line 42). Therefore, the overall extent of errors on verbal inflection in past tense is very severe.

What are the main patterns of the errors you identified? Are they systematic or variable? 

I can identify three main patterns of errors. Firstly, when the waitress used the verbs in simple past tense, she used the uninflected verb form instead of v-ed form, like ‘transfer’ in ‘Then I transfer myself to Tomlin’ (line 11) and ‘get’ in ‘Last year, I get no holiday’ (line 125); secondly, when she used the ‘be’ verbs, she used ‘is’ in ‘there be’ structure (line 2, 3, 66) and used ‘is’ as a linking verb (line 9, 12, 65, 111) and an auxiliary verb (line 9); thirdly, the incomplete use of present perfect tense (line 28, 43, 104, 124). There is one pattern that is rarer and not typical, thereby it cannot be judged on systematicity (her use of v-ing to substitute v-ed (line 42, 44)).

The first and second type of errors are systematic but the third type of error is variable. Firstly, with regard to the first pattern, the errors are very consistent with all 14 usages being infinitive verbs rather than verbs in the past tense. Secondly, in respect to the ‘be’ verb, all wrong usages are quite consistent. ‘Is’ is used incorrectly rather than any other ‘be’ verbs, except the usage of ‘am’ in ‘when I am schooling’ (line 42). There is only one correct usage of the ‘be’ verb in past tense, which is ‘was’ in ‘I go Chinese church when I was small’ (line 49). However, the correct rate is only 1/9, and out of the 8 wrong usages, 7 are consistent. Thirdly, in present perfect tense, the use is rather variable, ranging from missing the auxiliary verb ‘have’ (line 28) to using simple present tense or simple past tense to replace present perfect tense (line 43, 124, 104) and using the infinitive verb instead of past participle (line 124). 

To what extent do you see the role of L1 transfer?

It is possible that L1 transfer has played a major role in the speech of the waitress in the data but there is some contradictory evidence. In the L1, verbs do not inflect for past tense. This property is in line with the English speech of the waitress. In the conversation, she used infinitive verbs rather than inflected verbs in 24 cases, such as ‘transfer’ rather than ‘transferred’ as mentioned above, ‘get’ in ‘we get lessons’ rather than ‘got’ (line 44), etc. Only once did she use the inflected verb form correctly, namely ‘was’ in ‘when I was small’ (line 49). Moreover, among the five usages of present perfect tense, three used the infinitive verb form (line 43, 124). This makes up 27/31 of the usages of uninflected verbs for past tense. Therefore this target property is quite consistent to L1. However, there is one more important feature present in L1 which is lacking in the L2 speech, which is that past tense is expressed with other expressions such as the adverb ‘yesterday’. In the L2 data, there are only three usages of these other expressions that can indicate past tense, including ‘when I was small’ (line 49), ‘two three month back’ (line 64) and ‘Last year’ (line 125). Past tense in all the other usages could only be guessed out by the contexts. Due to this low rate in using necessary past-tense-indicative expressions, which is at odds with the L1 property, it is unconvincing to conclude that L1 transfer has played a major part in the waitress’ English. Thus, my answer is that this transfer is possible unless convincing data is given.

What further data you would like to see in support for the points you raised along the points above.

To begin with, since the waitress could speak several languages, namely L1 Teochew, Hokkien and English, it is hard to deem that only her L1 Teochew influenced her English. It is possible that her Hokkien interacted with her English as well. To determine the source of such transfer, data relating to if Hokkien features uninflected verbs in past tense and if it uses other expressions to show past tense should be given. If Hokkien does not have such a property, the probability of L1 transfer is higher.

Furthermore, due to the aforementioned contradiction, data regarding whether L1 Teochew often omits expressions that are indicative of past tense should be further provided. If such omission is frequent, then the degree of L1 transfer will probably be larger.

Lastly, in order to decide the generalisability of this L1 transfer, research should be conducted on three comparative groups of people: (1) people with the same L1 Teochew and same L2 English, (2) people with same L1 Teochew but different L2, and (3) people with different L1 but same L2 English. This is to investigate whether the transfer is due to L1 Teochew transfer or because of the development of L2 English. If the data from people in group 3 does not show the target property and data of all or almost all people in group 1 and group 2 show this property, we can conclude that the L1 Teochew transfer of this property is at a high degree.

Part II : Essay- Native Language Plays a Major Role in Second Language Acquisition.

Introduction

As a ubiquitous phenomenon, native language, namely L1, is acquired first before any other language, typically a second language. It has long been suggested that the acquisition of L2 is influenced by knowledge and constructs of L1 (Wang et al 2003). Schimke (2018) comments on this influence that one crucial feature of second language acquisition is that learners use their L1 knowledge to inform L2 learning. This influence can also be termed as transfer, which is alternatively defined as the precedingly acquired language skills to facilitate L2 acquisition (O’Malley and Chamot 1990). However, L1 transfer can not only be positive, but also negative. This distinction is mentioned in a large amount of literature (Chen 2020; Wang 2010;Vâlcea 2020). It seems natural that L1 plays a role in second language acquisition. Nevertheless, whether this role is major or minor needs to be further investigated. In my view, L1 plays a major role in the acquisition of L2. In this essay, I will demonstrate and justify this opinion, specifically by drawing upon two large-scale studies.

Review One and Discussion

To effectively justify my stance, the studies I review should be quantitative and they should involve a sufficiently large number of subjects. Meanwhile, the transfer should not only be verified but also be comprehended regarding the extent.

Participants and Background

The first review I chose is of a study (Siu and Ho 2015) based on 413 primary school students in Hong Kong whose L1 is Cantonese, a Chinese dialect, and L2 is English. Chinese is a language with scant inflectional variations whereas English is an inflection-abundant language. To give an example, the equivalent Chinese of ‘He wants to eat big meals’ is ‘Ta xiang chi da can’ (He want eat big meal). In the English sentence, ‘s’ in ‘wants’ and ‘s’ in ‘meals’ are both inflections while there is no inflection in the Chinese sentence.

In respect to the subjects, all participants were 6-to-10 years old students who had Cantonese-English bilingual education. In their daily life, Cantonese was used both in school and outside of school whilst English was only spoken in school. This is conducive to the probing of L1 transfer, as these students’ Chinese is more advanced than their English, thereby exhibiting a marked difference between L1 and L2.  

Test Design

The test item in this study is syntax, which can be considered as the morphological agreement, and the order of lexical items that make good sense. There were three tasks to test the students’ syntactic skills (SS), namely Oral Cloze (sentences with words missing), Morphosyntactic Correction (MC: sentences with grammatical mistakes) and Word Order Correction (WOC: jumbled words needed to be put in order). Note that word order knowledge is different from morphosyntactic knowledge. The central point of WC is ‘order’, and the focus of the MC is morphological agreement among the lexical items in a given text.

The task design for Chinese is that subjects were asked to match illustrations to 20 lexical items to measure their general oral Chinese proficiency, correct 12 jumbled Chinese sentences opted from local Chinese textbooks to measure word order skill, correct 15 sentences with classifier and preposition errors alike to measure morphosyntactic skill, carry out word reading, sentence comprehension and passage comprehension to measure literacy. Measurement of English was carried out in a similar manner, only that the test properties of English morphosyntactic skill were more numerous, which is tacitly reasonable owing to the more diverse English inflection system.

Data Analysis

After the tasks were conducted, the data was analysed. Correlation tests validated that Chinese RC positively correlated with Chinese SS, and the same pattern was found in English RC and English SS. Meanwhile, RC and SS presented strong relationships across L1 and L2.

To test the predictability or explanatory power of the SS within each language, Hierarchical Regression Analysis was performed with the condition that ‘age, nonverbal intelligence, working memory, oral vocabulary and word reading were statistically controlled’ (p.330-331). The result was that SS has significant association with RC in both languages. To be specific, regarding Chinese, the two SS in first grade had 9% explanatory power over RC, with the figure for WOC higher than MC. The two SS’s explanatory power over RC in third grade was 7%, with WOC and MC displaying relatively equal numbers. This is distinct from the results from English. In first grade, albeit a combined explanatory power of 4%, the morphosyntactic skill’s explanatory power over RC was undependable, while the figure for word order skill was reliable. In third grade, 9% of change in RC could be accounted for by the two SS, with MC exceeding WOC.

This has substantiated that SS could statistically explain RC within each language, but it was just the first stage before the more direct test of transfer across languages.

To test the transfer, Structural Equation Model was used. A direct effect model testing the effects of L1 SS on L2 RC was conducted, finding that ‘the direct path coefficients from L1 Syntactic skills to L2 reading comprehension was significant at ß = 0.50, p = .002, and at ß = 0.67, p < .001 for first and third graders, respectively’ (p.326). This result laid the groundwork and qualified the subsequent partial and full mediation tests and comparison.

In the following partially mediated structural models (excellent fit), the predictive effects of L1 SS over L2 SS was significant (‘ß = 0.65, p < .001’ and ‘ß = 0.66, p < .001’ in first grade and third grade respectively (p. 326)) , and L2 SS had comparably significant predictive effects on L2 RC (‘ß = 0.78, p < .001’ and ‘ß = 0.99, p < .001’ in first grade and third grade respectively). Meanwhile, L1 SS significantly indirectly predicted L2 RC through the medium of L2 SS (‘ß = 0.51, p < .001’ in first grade and ‘ß = 0.65, p < .001’ in third grade). However, without the medium of L2 SS, these two factors lost the direct effects. 

The ensuing analysis demonstrated that L1 RC did not have the explanatory power over L2 RC (ps > .05 in both direct and indirect effects). This suggested that L2 SS had mediating effects on L1 SS predicting L2 RC, but L1 RC did not.

In the fully mediated model, there was an excellent fit. Comparing the two models with chi-square difference, which was not significant, the fully mediated model was favored. L2 SS showed full mediating effects on L1 SS and L2 RC (mediation effect at 0.5, p < .001 for first grade and that at 0.65, p < .001 for third grade).

So far, this experiment has proved L1 transfer on L2, specifically L1 transfer of syntactic skills, but it is more comprehensive to test each individual syntactic skill. Thus, a likewise analysis was further conducted, exploring the relationship between L1 WOS and L2 RC through L2 WOS, and the same for morphosyntactic skill. Regarding the first exploration, both grades showed significant direct effects at 0.32 and 0.37. However, when comparing the fit of the data, partially mediated models were superior to the fully mediated models. In both grades, the explanatory power of L1 WOS over L2 WOS, L2 WOS over L2 RC and L1 WOS over L2 RC were all significant (p < .05) and L1 WOS showed both direct and indirect effects on L2 RC (ps < .05). The result was similar in the second exploration, except that the fully mediated model was chosen for first grade. Therefore, the data of individual syntactic skills were consistent with that of them combined. 

Summary of Review One

This study corroborated that there is significant transfer of L1 SS to L2 RC through L2 SS. However, L1 RC failed to mediate the above transfer, indicating that the similarity between L1 and L2 was employed to aid L2 learning. Given that Chinese and English are more typologically distinct, there should be more and stronger transfer in less typologically distinct languages. Even though I do not generalize the transfer, which would entail drawing on all possible knowledge available in L1, L1 syntactic skills alone can have powerful transfer on L2 in that syntactic skills significantly affect reading comprehension (Barnett 1986). Taking into account how long a typical learner engages in L2 learning without a L2 living context — it takes up to 5 years and 7 years for English learners in immersive English living contexts to achieve oral and academic proficiency respectively (Hakuta and others 2000) — it is apparent that native language plays a major role in Second Language Acquisition. 

Review Two and Discussion

Participants and Background

The last review has a focus on L1 transfer based on correct points students won, without much analysis of their errors. In comparison, this second review of a study (Chan 2004) is based on the same contexts as the last one but with a focus on negative transfer errors and with more engagement of L2 English learners of disparate levels, including more advanced learners such as college level. 710 students (486 lower-intermediate students and 224 upper-intermediate students) were selected as subjects.

There are five common types of Chinese transfer errors of English, namely copula errors (Chinese often drops copula), placement of adverb errors (adverbs are often preverbal in Chinese), existential structure (Chinese use a word meaning ‘have’ in English), relative clause errors (Chinese head nouns are head-final and do not use any relative pronouns and adverbs), and verb transitivity (some English transitive verbs are intransitive in Chinese, and vice versa). In this essay, these errors are referred to in the above order.

Test Design

The first process used interviews of 42 students. One part of the interview is that they tried to translate pre-determined Chinese into English and recall the extent to which they thought via their L1 Chinese. The second part required the volunteers to translate Chinese into English with aid and instructions from a Cantonese speaker (specifically, volunteers are given word cards to choose from). 

The second process engaged all students in tasks consisting of translation of ten pre-determined Chinese sentences with two of each type of error, and grammar judgement comprising 15 grammatically incorrect sentences, 3 correct sentences without any of the five types of errors and 2 correct sentences without relevance. 

Data Analysis

The data from the interviews displayed that about 75% of both the upper-intermediate and lower-intermediate had a strong tendency to use Chinese to think. In the first part, though most students translated the given Chinese holistically instead of translating word-by-word, the corresponding transfer error rates were 38%, 13%, 56%, 38% and 75%, indicating a strong transfer derived from Chinese structures. The instructed translation task in the second part witnessed a notable drop in such error rates, except relative clause errors remaining relatively the same.  

From this interview data alone, we can speculate that students drew on some features of Chinese to formulate English sentences passively or actively and this suggests Chinese transfer has played a role in their English production. Even though it is insufficient to generalise these speculations based on 42 participants, it seems to prove that my stance is preliminarily sound. 

The data from translation tasks in the second process was summarised in the table below (table 1). All errors were above 45% except Incorrect Placement of Adverbs, and students of high proficiency were better in this aspect, making less errors, but the errors were still significant. This validated that regardless of English proficiency, a high proportion of students employed L1 knowledge, and the produced English sentences had a very notable element of Chinese regarding syntactic structures.

In the grammar judgement, a high proportion of students failed to recognise the errors in the incorrect sentences (74%, 53%, 70%, 67% and 82% for the lower-Intermediate students, 39%, 13%, 29%, 30% and 68% for Upper-Intermediate students), and a third of students misjudged the correct sentences. This, again, proved the verdict in the translation tasks.

Summary of Review Two

The data from the interviews revealed the L1 transfer, owing to the use of Chinese knowledge in English production. The data from translation tasks and grammar judgment tasks, with sufficient subjects, further generalise this L1 transfer, and even more advanced English learners could not evade this influence. In short, the data supports the assertion that L1 transfer was significant and higher L2 proficiency can not fully counteract this transfer. 

Conclusion

Both reviews confirm that L1 can actively and passively influence L2. The first study attested to L1 syntactic transfer, which is ample evidence that native language plays a major role in second language acquisition due to the significant role syntactic skills play in languages. The second study proved that errors resulting from L1 characteristics are extremely common, testifying that L1 transfer is ubiquitous and even high L2 proficiency with many years of study cannot steer clear of this. Indeed, high L2 proficiency comes at a great time cost and energy cost, dooming most L2 learners at low-proficiency levels, which amplifies L1 transfer’s power. The conclusion is that native language plays a major role in second language acquisition.

References:

Barnett, M.A. (1986) ‘Syntactic and Lexical/Semantic Skill in Foreign Language Reading: Importance and Interaction’ The Modern Language Journal, 70(4): 343-349 <https://www-jstor-org.sheffield.idm.oclc.org/stable/326811?sid=primo&seq=1#metadata_info_tab_contents>[accessed 26 Jan 2021]

Chan, A.Y. (2004) ‘Syntactic Transfer: Evidence from the Interlanguage of Hong Kong Chinese ESL Learners’ The Modern Language Journal, 88(1): 56–74 <https://www-jstor-org.sheffield.idm.oclc.org/stable/3588718?sid=primo&seq=1#metadata_info_tab_contents>[accessed 26 Jan 2021]

Chen, C. (2020) ‘A Study on Positive Transfer of Native Language and Second Language Teaching Methods’ Theory and practice in language studies, 10(3): 306–312 <https://search-proquest-com.sheffield.idm.oclc.org/docview/2382073908?https://search.proquest.com/pq1lit&pq-origsite=summon>[accessed 26 Jan 2021]

Hakuta, K., Y.G.  Butler & D. Witt, (2000) How Long Does It Take English Learners To Attain Proficiency? University of California <https://www.researchgate.net/publication/2436759_How_Long_Does_It_Take_English_Learners_to_Attain_Proficiency>[accessed 25 Jan 2021]

O’Malley, J. M. & A. U. Chamot (1990) Learning Strategies in Second Language and Acquisition, New York: Cambridge University Press.

Schimke, S. et al. (2018) ‘First Language Influence on Second Language Offline and Online Ambiguous Pronoun Resolution’ Language learning, 68(3): 744–779 <https://doi-org.sheffield.idm.oclc.org/10.1111/lang.12293>[accessed 25 Jan 2021]

Siu, C. T. & C. S. Ho (2015) ‘Cross-Language Transfer of Syntactic Skills and Reading Comprehension Among Young Cantonese-English Bilingual Students’, Reading Research Quarterly, 50(3): 313–336 <https://search-proquest-com.sheffield.idm.oclc.org/docview/1695779610?pq-origsite=summon&https://search.proquest.com/pq1lit>[accessed 26 Jan 2021]

Vâlcea, C.S. (2020) ‘First language Transfer in Second Language Acquisition as a Cause for Error-making in Translations’, Diacronia, 11: 1-10

Wang, M., K. Koda, & C. A. Perfetti. (2003) ‘Alphabetic and Nonalphabetic L1 Effects in English Word Identification: A Comparison of Korean and Chinese English L2 Learners’, Cognition, 87(2): 129–149 <https://doi.org/10.1016/s0010-0277(02)00232-9>[accessed 22 Jan 2021]

Wang, Y.Y. (2010) ‘Negative Transfer and Acquisition of Non-agentive Passive Voice’. Studies in Literature and Language, 1(7): 76-81