Review of C. Hill & K. Parry (Eds.), From Testing to assessment: English as an international language.

By Tom Cobb, École de langues, Université du Québec à Montréal, Montréal, Canada.

Instructors and course designers who construct reading tests for students of English will find here a fascinating context in which to reflect upon their endeavours. They may be surprised to learn that English literacy testing plays a larger and more controversial role in the developing world than they had realised, for example helping maintain social divides across whole nations (Nigeria, Zimbabwe). They will consider the possibility that the very idea of literacy testing is a legacy of the Mandarin Chinese, imported into Europe in the 18th century when literacy provided a less visible basis than wealth and privilege on which to re-anchor shifting social hierarchies. They will glimpse some of the subtle errors that can creep into the best intentioned literacy testing, for example when adapting a children's story for adult literacy testing leaves in place an underlying schema of childhood, leading adult testees into predictable errors and misdiagnosis of the true state of their literacy. Conversely, they may be impressed by the awesome amount of labour that goes into equalizing comprehension questions between versions of a TOEFL reading test.

These are some of the themes brought together in this fine collection. Yet the pieces are not a random sample of views, but fit (mainly) within a theory of what reading tests are and what is wrong with them. In the editors' overview sections, literacy tests are seen in a negative light. These tests wreak untold damage on the world, not because of errors, cynicism, or ignorance in particular cases, but because they derive from an unacceptable model of literacy, a Western model most clearly articulated by Olson (1977). Olson argued that the nature of literacy was best understood in contrast to orality. In an oral culture, whether that of childhood or the various oral and semi- oral cultures that remain in the world, information is rarely conveyed by texts or even speech alone, but as much by surrounding context and situation, with strong contributions from participants' expectations and prior knowledge. But in a literate culture, a written text is more autonomous, abstracted from any particular context or situation, and much less dependent on expectation and prior knowledge. Hill and Parry rightly argue that this version of literacy is not the only legitimate one, and certainly not the one that prevails in many sub-cultures at home or main cultures abroad, but is nonetheless the version the tests test.

Their solution is to reconceive the nature of literacy and literacy testing. If reading within a learner's culture normally takes place within a rich context of familiar and shared information sources, then reading tests of English should be designed within a similar framework. One piece in the collection discusses an attempt in Zimbabwe to implement this notion. Reading tests were constructed not from abstract texts the testees had never seen on subjects they had never heard of, but from an assortment of local newspaper clippings, TV schedules, and bank deposit forms. Another piece describes an attempt in Australia to test students via the portfolios they had put together in groups, reflecting the naturally social nature of reading and writing.

Unfortunately, the effort to house the collection under a single roof is also its weakness. It involves a distortion of Olson's position, a failure to mention what Olson sees as the advantages of extending literacy beyond orality, and occasionally leads the editors into absurd positions. First, Olson (1977) never claimed that Western literacy ever managed to produce totally free-standing texts, demanding no contribution whatever from the reader, so Hill and Parry's revelation that even educated readers can hardly read some science texts for which they have inadequate prior knowledge misses the point. Olson merely argued that the evolution of text in the West has been toward self-containment, not that it ever could or should arrive there. The advantages of a degree of text autonomy are, in summary, that it releases individual thought from group-think; that it allows for the encoding of genuinely novel ideas; that these ideas can be comprehended by others far away in space, time and situation; and most importantly that it allows writers to use their own texts as sources of learning. Whether true or false, these claims deserve a hearing.

From my perspective of a decade teaching and testing Arab university students in English reading, I know how badly literacy testing can be done, how iniquitous its function can be, and how easily one ends up testing students on things they were never taught. I am also convinced that the model of literacy my students have grown up with, with its heavy situational-social-oral component, is valid and has important functions for them. However, I strongly believe, and I think my students do as well, that they need to expand their repertoire of literacies to include the comprehension of texts that are not part of their tradition, texts which often contain unexpected information, sometimes about their own culture and how it is seen by others.

For example, texts dealing with the conditions of democracy in Saudi Arabia, or female circumcision in the Horn of Africa, are unlikely to emerge from the rich social discourse these students have grown up with. But well reasoned and researched texts on such subjects are more and more accessible in countries like Saudi Arabia and the Sudan, less and less affected by the censor's felt pen and scissors, thanks to electronic media. The only question now is whether anyone there has been empowered to read them. These texts require readers to go beyond the constraints of familiarity, expectation, and prior knowledge if they wish to respond to the points their authors are actually making. Getting students in developing countries to read widely in a range of text types, and giving them tools for reading the less familiar types, is the literacy challenge their instructors face--not choosing between models of literacy on their behalf.

As mentioned, the attempt to lay all problems at the door of "Western literacy" leads the editors into occasional absurdities. One of the pieces tells of an interview in Canada between a functionally illiterate man and a literacy assessor. The assessor asks the man to extract information from familiar texts like store flyers and road maps. "Which of these meats is cheapest?" (from four choices with words, pictures, and prices), and "How would you get from here to Toronto?" (on a map with one road marked "To Toronto.") He answers both questions correctly. Asked to explain his information processing, he says he recognised the picture of the cheapest cut because he always buys it, and could produce the road number because he often drives it. Not surprisingly, in her report of the interview the assessor refers to such heavily context-driven "reading strategies" as weaknesses. For this, Hill and Parry criticise the assessor; she has failed to value "the student's use of real- world knowledge" and "focuses instead on low-level problems" (such as failing to decode place names). This they trace to her "lingering commitment to an autonomous model of literacy" (p. 264).

Some night, lost in a blizzard 100 miles out of Sault-Sainte-Marie, the man will thank whoever took the time to help him over the "low-level problem" of not recognizing what the symbols mean on his map. The thing about maps is, you use them when your prior knowledge is not enough -- when you are in a place where you have not been before.

Olson, D. R. (1977). From utterance to text: The bias of language in speech and writing. Harvard Educational Review, 47 (3), 257-281.

Tom Cobb
City University, Hong Kong
January 1997