Discussion of the Translation Session in Shanghai Gaokao English Examination from a Critical Perspective

丁磊

1. Introduction

This paper aims to discuss and critique the translation session in Shanghai Gaokao English paper. Translation here refers to translating a Chinese sentence into corresponding English versions and students are required to use given key words or expressions in their translation. For example，

雖然現代社會物資豐富，給予消費者更多的選擇，但也使不少人變成購物狂。（turn）

Though abundant supplies give more choices for the customers in modern society， they also turn many people into a shopaholic.

（Shanghai Municipal Educational Examinations Authority，2014）

Translation has been adopted as a test format since 1998 and the number of sentences to be translated is four from the year 2017. Not only does the translation format test the basic skills of translation， but also it tests the ability to comprehensively use vocabulary， grammar and sentence structures and furthermore， it involves students proficiency in differentiating the difference between Chinese and English， especially when dealing with the translation of Chinese proverbs （Shanghai Municipal Educational Examinations Authority，2015）.

Therefore， it is not hard for us to realize that the effectiveness of translation as a test can exert significant influence in English teaching and learning because as long as a test is considered important， like Goakao， the preparation for it is prone to dominate and guide all teaching and learning activities （Hughes， 2003）. However， the research into translation as a test format on the secondary school level is rare， which may be attributed to the fact that it seems only Shanghai tests translation in Gaokao so far， which is worth critical discussion， though.

2. Discussion

2.1 Reliability

In any translation test， reliability is the first and significant aspect to be taken into account and guaranteed. （Li， 2013）. Reliability concerns the ‘consistency of test judgments and results （Davies， 1990， p. 21） and according to Hughes （2003）， the reliability of a test can be quantified by analyzing its reliability coefficient， which enables us to compare the reliability of different tests. In other words， if a reliability coefficient of a test is 1， to great extent， it shows that the test can offer identical results for a particular group of candidates in spite of when it happened to be carried out （Hughes， 2003）. Since the overall scores of the translation session have varied a bit， the passing rate is used to compare in the present study and the passing rate of the translation session seems quite stable （around 0.6）， therefore in this case， the current translation test appears reliable.

However， it is not convincing to judge the reliability of a test only from one perspective， so Li （2001） asserts that the reliability of a translation test can be judged from the quality and quantity of the test， the implementation of the test and assessment of the test.

2.1.1 Test contents

It is acknowledged that the test contents should be consistent to the language skills to be tested （Li， 2013）. Unfortunately， it is true that the current translation session is to test students ability to translate correct and fluent sentences by using the language knowledge they have acquired， which is clarified by Shanghai Municipal Educational Examinations Authority （2015）， but the test contents lack variety as only sentences are required to be translated. Also， sentence translation restricts students in bringing their translation proficiency into play （Li， 2013）. Since translation is to test students compressive use of language as mentioned above， accordingly， the test contents need diversity as well， which， to some extent， can ensure the objectivity and consistency in the assessment of the test （Zou， 2005）. Based on the students language proficiency and practicability of the test， it is suggested to add short paragraph translation here because any language teaching and learning activities cannot separate from the context. As Shao （2003） points out that translating a passage can better reflect students ability to use language and to tell the difference between Chinese and English in terms of semantics， sentence meaning， rhetoric and cultural understanding. In short， in terms of test contents， current translation testing seems not that reliable.

2.1.2 Test assessment

The reliability of the test assessment involves the test-raters consistency and inter-raters consistency. In other words， every test-rater has to follow the assessment criteria in a constant and strict manner （Li， 2013）. Translation， as an objective testing， can be assessed through mechanical method， impression method and analytical method （Liu， 1991） and the mechanical method is used to assess the current translation testing， in which marks are reduced related to the quantity of errors in the translation. Lets look at the translation session in 2014 as an example. The score for 5 sentences is 4， 4， 4， 5， 5， respectively. Test raters will be classified into several groups and each group will mark specified one or two sentences. Furthermore， each group has a leader， who is responsible for setting the detailed assessment criteria based on the reference answers provided by test-makers， and before setting the detailed criteria， test-raters and the leader randomly pick up several papers to mark and discuss. Also， every sentence is marked by two raters on the computer， which is delivered automatically and confidentially and the mean score of the two raters is the final score for the student. However， if the score gap between the two markers is over 1， this sentence will be delivered to the group leader to mark again. Although this kind of marking is considered to reflect atomistic view of language， which separates the language patterns unnaturally （Li， 2001）， considering the testing contents and stake of Gaokao， the present assessment seems to be fair and reliable.

It is true that after repeated practices， most of students manage to translate sentences quite fluently and correctly but the elegance of the translation fails to show up and the score lacks discrimination if applying mechanical method （Li， 2013）. Therefore， it is advised to combine synthetic approach and analytic approach to assess （Li， 2001）. But it is impossible to assess only one test pattern by using different assessment approaches and thus test contents need variety. To sum up， test contents and assessment are indispensible and intertwined to each other.

2.2 Validity

Language testing is an approach to measure ones language proficiency in different situations and for different purposes （Allison， 1990）. That is to say， if the score of a test can indicate ones language proficiency， that score has to be reliable and valid as well. Hence， reliability and validity are interlocked as reliability is not a sufficient but essential component of validity （ibid.） Also， Alderson， Clapham & Wall （1995） claims that a test is invalid if it is not reliable. Validity was described by Davies （1990） as ‘the truth of the test and ‘its relation to what is intended to test （p. 21）. In this paper， the validity of the translation session will be discussed in view of content validity and criterion-related validity.

2.2.1 Content validity

As Li （2013） puts forward， the validity of a test can be measured through the relationship between the test contents and the abilities to be tested. Since the translation session is to test students comprehensive ability of using language as mentioned above， the language ability to be tested can be detailed and classified as usages & sentence patterns， grammatical patterns and proverbs. If we measure the validity in terms of contents consistency， the translation testing seems stable and valid in that it is quite clear that sentence patterns and expressions are put emphasis on， which accords with the testing goals—students are expected to translate fluent and correct sentences by making use of language knowledge they have learnt.

Nevertheless， as a productive skill， translation cannot remain stagnant only on the level of words or sentence patterns， because it is easy for students to achieve it through rote learning and repeated practice and as a result， translation turns into mechanically putting separate word chunks or patterns together， let alone the meaning and culture. On the other hand， it is not easy for universities to select students with outstanding translation proficiency because the contents are too easy to tell from one and another student. Therefore， in order to well achieve the validity in this case， variety and difficulty of the contents need to be promoted.

2.2.2 Criterion-related validity

There are two types of criterion-related validity， defined by Hughes （2003） as concurrent validity and predictive validity. The former is set up when the test and the criterion are carried out at about the same time while the latter concerns the extent to which a test can predict an individuals performance in the future （Hughes， 2003）. So the validity of the present translation session is discussed in terms of concurrent validity.

Unfortunately， it seems impossible to apply criterion-related validity under the current situation because there is not a specific time for translation. The whole English test in Gaokao lasts for 120 minutes and the listening session starts at first and lasts for 18 minutes， after which students can decide to finish any session freely. Let us consider this situation. Student A has a distinguished ability in translation but he finished translation sessions only in 5 minutes， for he was stuck in other sessions while student B is ordinary in translation but he spent 10 minutes in doing translation and checking errors. Consequently， student B got higher marks than student A in translation session， and the question then arises： is it a valid measure for students translation ability？ In fact， this question applies to other sessions like reading and writing. We cannot say it is definitely unfair for student A as he didnt manage time well but the lack of specified time for testing a particular skill， to some extent， does fail to promise the validity since students language ability can be judged from his speed and accuracy in the test （Li， 2013）.

To sum up， it is proposed to set a fixed time for each session in Gaokao based on the test contents and practicability.

2.3 Backwash

Backwash is defined by Hughes as ‘the effect of testing on teaching and learning （2013， I） and he even points out that any language testing will generate harmful or beneficial backwash， so will the translation testing. The present translation session in Gaokao boasts both beneficial and harmful backwash.

2.3.1 Beneficial backwash

Compared with non-translation testing， translation can facilitate the improvement of students English spelling ability because students attach importance to the words spelling and memorization when they learn new words， otherwise they may fail in the translation session， which consolidates the learning results （Qin， 2009）. Also， translation helps to better reflect students acquisition of vocabulary and grammar from a micro perspective （ibid.） and furthermore， since a target word is required to use， the testing targets are clear and as a result， teaching and learning activities can be more focused in order to optimize the effectiveness and efficiency. Last but not least， since writing is comprised of different sentences， a good translation skill seems to be conducive for the accuracy and fluency of writing.

2.3.2 Harmful backwash

Since translation is assessed by mechanical method， which only pays attention to the accuracy of the translation， teachers and students are likely only to pursue the achievement of accuracy， ignoring the whole sentence. As a result， some translation versions are correct but read unnatural， for they translated words for words and whats more， if students learn and practice English in that way， their English learning will be deeply impacted by Chinese ， which will absolutely block their further improvement of English. For example，

不在房間的時候別開燈。

Dont leave the light on when you are not in the room. （√）

However， some students translated it as ‘Please dont turn on the lights. as they failed to understand the sentence from a whole picture and it is apparent to find that they think about the translation from Chinese perspective rather than English. In addition， some teachers even offered very bad translation contents to students， just to allow students to practice some target words， which can be seen as a disaster for teaching and learning. Students are apt to get bored with repeatedly translating long， monotonous or meaningless sentence so as to get good marks in translation testing， which greatly goes against the original intension of translation exercises—to help students comprehensively practice and use the language.

3. Conclusion

To conclude， if the current translation session can add variety and authenticity of test contents， apply different assessment approaches and set a specified testing time， it can be a more reliable and valid testing format. Also， testing cannot dominate teaching and learning but testing can serve as a channel to facilitate and reflect on teaching and learning. Teachers are advised to make use of different ways to teach and help students to practice translation while students themselves had better take translation as a stage on which you can bring the language knowledge you have acquired into full play. Therefore， more beneficial backwash can be generated. Although it seems impossible to make a perfect testing especially when you take into account reliability， validity， backwash， practicability， stake， and so forth， but we still can try our best to polish the testing.

References：

[1]Alderson，J.C.，Clapham，C.&Wall，D.（1995）.Language test construction and evaluation.Cambridge：Cambridge University Press.

[2]Allison，D.（1999）.Language testing & evaluation： An introductory course.Singapore： Singapore University Press.

[3]Davies，A.（1990）.Principles of language testing.UK： Basil Blackwell.

[4]Liu，R.Q.（1991）.Language testing and its methods.Beijing： Foreign Language Teaching and Research Press.

[5]Li，X.J.（2001）.Science and art of language testing.Changsha：Hunan Education Press.

[6]Li，J.C.（2013）.Translation competence and assessment of reliability and validity in the translation testing.Journal of Xian International Studies University 21（2）： 117-121.

[7]Hughes，A.（2003）.Testing for language teachers（Second edition）.Cambridge： Cambridge University Press.

[8]Qin，Y.（2009）.Backwash and its problems in translation testing.Journal of Inner Mongolia University for Nationalities 35（5）： 68-70.

[9]Shao，Z.H.（2003）.Comparison between Chinese and English and translation —evaluation for translation session in TEM8（2002）.Chinese Translation Journal（1），76-78.

[10]Shanghai Municipal Educational Examinations Authority.（2014）.The evaluation for College Entrance Examination Paper for English（2014）.http：//www.dx513.com/tzcz/xyjs/sh/201406/767904.html.

[11]Shanghai Municipal Educational Examinations Authority.（2015）.Shanghai Gaokao workbook.Shanghai：Shanghai Guji Press.

[12]Zou，S.（2005）.Language testing.Shanghai： Shanghai Foreign Language Education Press.