【摘要】：Long before scientific language testing cast its first light, teachers had been adopting translation on a daily basis to test students’ language proficiency. Till very recently, language tests incorporating translation have been mushrooming in China. Thus, a picture of old and new tests of translation unfolded before us. Some aim at English reading proficiency with English-Chinese translation, some seek to test language proficiency by two-way translation, others concentrate on translation or interpretation proficiency in a way that translation is their sole or major constituent.
No matter what their purposes are, an authoritative voice challenging or objecting to these practices has been echoing in our mind. In the 1960s when structuralism was popular, Lado (1961:33) raised questions against the test of translation in language tests. In the 1990s when China was gaining speed in the study of language testing, Song Zhiping (1997:30) and Xu Lina (1998:29) pointed out problems in existing tests of translation for translation’s sake.
The polar difference between theory and practice has put many testers and teachers in a dilemma: should we use tests of translation? And those who are interested in language testing would like to ask, “Why is there such a difference?” Stevenson (1982:153) identified the root causes by concluding that “the basic problem with translation testing is therefore first, our traditional inattention to testing considerations and only secondarily, that delightfully messy area we term translation theory and practice.” Despite its enlightenment, this conclusion seems too broad to be out-of-box ideas. Instead, we found breakthroughs by examining the reasons listed by another objector (Heaton 2000:18). He pronounced that “tests of translation tend to be unreliable” because of the complex nature of the various skills involved, the methods of scoring, the highly artificial sentences and literary texts for
translation and testers’ unrealistic expectations of examinees.
All these challenges, on the other hand, functioned like the voice on the road to Damascus. We were inspired with ideas to construct a well-defined model of translation competence, write various translation tasks and improve the scoring with construct validity, reliability, authenticity and practicality in mind. In this way we propose the following research questions:
1) What is the model of translation competence for English majors? Or what are the sub-skills of the translation competence?
2) How can we enrich translation task types?
3) How can we improve the scoring of subjective translation tasks?
Answering any of these questions would call for an organization’s long-term investigation and research, thus the author chose to present a preliminary study in an attempt to stimulate further research and refinement.
Having realized that translation testing needs to be considered from both testing and translation perspectives, we now look at the first angle.
By reviewing the development of language testing, we found four major approaches towards English testing. The first one is the essay-translation approach in the pre-scientific era; the second is the structuralist approach with multiple-choice questions as its representatives; the third is the integrative approach best reflected in cloze and dictation; and the fourth is the communicative approach which puts authenticity and communicativeness on its banner and billboard. Although we stated the four approaches in a time sequence, we observed that they were not confined to a specific period in history and each had its upsides and downsides. Therefore, a good test never refuses the synergy of more than one approach. We admit, however, that the communicative approach is growing in popularity and we now examine three aspects of the latest approach.
Our first focus is on the model put forward by Bachman (1990:84-98) of communicative language ability. It mainly consists of language competence, strategic competence and psychophysiological mechanisms. This comprehens