BACHMAN， LYLE. F. Fundamental Considerations in Language Testing. Oxford： Oxford university Press， 1990. Pp. xi， 408. ?12.95， paper. ISBN 0-19-437003-8.

Bachman （1990） writes about theoretical and practical considerations of language testing in this book which is recommended as a “must read” academic work for those serious students of language testing. The book is presented as seven chapters： measurement， uses of language， communicative language ability （CLA）， test methods， reliability， validation， and some persistent problems and future directions.

One of the most important concept in the book is the framework Communicative Language Ability （CLA） proposed for better understanding the relationship between language skills and language acquisition （Bachman， 1990）. Three components are presented in this framework： language competence， strategic competence， and psychophysiological mechanisms. Another important concept in this book is validity. Before Bachman， the validity of test interpretations was presented as several types such as content validity， criterion validity， and construct validity. However， Bachman （1990） emphasizes that “validity is a unitary concept related to the adequacy and appropriateness of the way we interpret and use test scores” （pp. 289）.

One can not agree more with this “unitary validity”. In China， validity is a kind of componential concept almost in all text book of language testing （Zou， 2005）. Zou （2005） introduces that from 1940s to 1980s， validity was divided into different types such as content validity， predict validity， and constructive validity， empirical validity， factorial validity， curricular validity， etc. However， Messick （quoted in Bachman， 1990， pp. 236） views that “validity is an integrated evaluative judgment of the degree to which empirical evidence and theoretical rationales support the adequacy and appropriateness of inferences and actions actions based on test scores.” What’s more， Bachman （1990） supports this view in Standards for Educational and Psychological（quoted in Bachman， 1990， pp. 237）. Validity will be meaningless if it is split into different parts （Zou， 2005）. For example， we can not implement a test without considering its construct validity even though it has qualified content validity. Only all types of validity are qualified， can test be meaningful and implementable. In addition， some of so-called types of validity connect with each other， in other words， they are indivisible. The construct validity of a test， to a large extent， depend on the content validity of the test. For example， items in a writing test can only be short writings or long writings but not cloze or multiple choice. As a result， compared with “componential concept”， Bachman’s （1990） “unitary concept” is more powerful and reasonable.

Maybe it is not agreeable that “tests through “real-life” approach cannot be used to make inferences about levels of language ability” （Bachman， 1990， pp. 356）. In the last chapter， Bachman （1990） introduces that the most complex issue is the authenticity which refers “the relationship between real life language use and the language use required by language tasks”. He demonstrates “real-life” and “non-test communicative” approach to describe the relationship between language use and test. However， because of the complexity of test takers’ needs of language use， he claims that the test through “real-life” approach can not indicate test takers’ language performance （Bachman， 1990， pp. 289）. Exactly this view may be controversial.

First， “real-life” approach means that we make a criteria which is “real-life” language use by which we design test tasks and infer test takers’ language ability （Bachman， 1990， pp. 289）. It means that the more test tasks look and operate like “real-life” language use， the more test scores can predict test takers’ language performance. For those whose language use needs are identical， this approach provides a helpful and practical test design； otherwise， this approach is quite difficult to be implemented （Bachman， 1990）. However， its complexity does not mean that we deny the relationship between “real-life” language use and language tests. We learn language for the purpose of communicating in real life which can be seen as language ability. Language tests aim to test our language performance and ability. Therefore， only in a “real-life” approach， our true language ability can be inferred with the least error.

Second， the emphasis of input in usage-based approaches to SLA can support this “real-life” approach as an indicator of language ability. In one hypothesis of the usage-based approaches， they claim that “exposure to input quantities and qualities typical of naturalistic-immersive contexts will engage procedural memory system optimally” （Van Patten and Williams， 2015， pp. 258）. We can easily find that the hypothesis emphasizes the importance of naturalistic characteristic， in other words， the “real-life” feature of language input. Therefore， “real-life” factor does plays a very significant role in our language use. It supports the view that “real-life” approach can estimate our language performance and ability.

Third， another support of “real-life” approach is the definition of “authentic test” and “authentic materials” in studies of language testing. Doye （quoted in Bailey， 2012， p. 269） defines authenticity as： “ An authentic test is therefore one that reproduces a real-life situation in order to examine the students’ ability to cope with it”. It means the inference of students’ true language ability depends on the production of “real-life” situation in language tests. In the literature review of Bailey’s （2012） study， the intuitive definition of “authentic materials” are presented as those materials in the “real-life” situation in foreign language. Therefore， there is no doubt that only in “real-life” language environment， can we measure the true communicative language ability. From this view， the importance of “real-life” situation in language tests is stressed again.

According to the analysis above， we have to admit that “real-life” approach is a powerful indicator of the language ability of test takers. However， just like Bachman （1990） mentioned that facing the complex variables in “real-life” language use approach， we can not implement this approach at present. Put another way， this complexity offers new directions and challenges to future research.

Generally speaking， this book has presented transparent language， sufficient knowledge and logical distribution. In addition to its academic value， this book is recommended for the presentation of all kinds of charts and simple language style. In spite of these outstanding characteristics， the presentation of few terms are not sufficient. For example， in chapter 6 Bachman （1990） claims that G-theory is an extension of CTS and it overcomes many of the limitations of CTS. However， CTS receives primary attention and be presented in a long passage while G-theory is introduced briefly only for ten pages. Because of the outstanding characteristic of G-theory， one would have liked to have seen a fuller treatment of G-theory.

In a nutshell， this book is notable and should be recommended to language testing majors worldwide. Just like Spolsky （2014） says， “This is a fine and original presentation of the state of that art in language testing.”

References

[1]L. F. Bachman. （1990）. Fundamental considerations in language testing. New York： Oxford University Press.

[2]Bernard. Spolsky. （1991）. Book review： Fundamental considerations in language testing. Oxford University Press. The Modern Journal. 75. 499-500.

[3]Bill VanPatten Jessica Williams. （2015）. Theories in second language acquisition： An introduction. New York and London： Routledge.

[4]K. M. Bailey. （1996）. Working for washback： A review of the washback concept in Language testing. Language Testing. 13， 257 - 278.

[5]Tim. McNamara. （2003）. Book review： Fundamental considerations in language testing. Oxford University Press， Language testing in practice： Designing and Developing useful language tests. Language Testing. 20， 466 - 473.

[6]Zou Shen. （2005）. Language Testing. Shanghai： Shanghai Foreign Language Education Press.

作者簡介：張寶欣，1993年9月28日出生，女，漢族，陜西咸陽市人，現(xiàn)就讀于西安外語大學(xué)英文學(xué)院研究生部2016級外國語言學(xué)及應(yīng)用語言學(xué)專業(yè)。主要研究方向：測試學(xué)。