2.3.3 Evaluation from a Cross-cultural Perspective

Despite a relative uniformity of academic texts imposed by requirements of the genre，speakers or writers of different mother tongues and cultural backgrounds may display different uses and text constructions.Different ways in which evaluation is conveyed are sometimes investigated between texts written by academics with different cultural backgrounds，but using the same language as a means of expression.Okamura（2005）analyzes the type and tense of verbs with we in British and Japanese scientists’research articles in English.The analysis reveals subtle differences between the British and Japanese papers，which indicates possible disadvantages for Japanese scientists.Valero-Garcés（1996）explores the cultural differences between texts written in English by Spanish-speaking academics and by Anglo-American academics with respect to metatext used in papers from economics journals.The results indicate that the Anglo-American writers use more metatext and are more concerned with guiding and orienting the reader.Scollon&Scollon（1995）report an Asian preference for strategies of indirectness arising from different cultural surroundings and participant roles in interaction.Likewise，Finns（Mauranen 1993；Ventola 1992），Japanese（Harder 1984；Hinds 1987），Malays（Ahmad 1995），Koreans（Choi 1988；Eggington 1987）and Chinese（Bloch and Chi 1995；Hinkel 1997）appear to favour a more cautious style of expressing opinions.

EFL（English as a Foreign Language）students（and even those who have a good control of English grammar and lexis）sometimes find it difficult to express evaluation in English，resulting in what Thomas（1983）calls“cross-cultural pragmatic failure”.Such contrastive studies are often conducted in the context of ESP teaching.For example，Hewings&Hewings（2002）consider the different ways in which interpersonal it-clauses are used in published journal articles and in student dissertations written by non-native speakers of English.Robberecht&Van Peteghem（1982）refer to the difficulties faced by Dutch and French students in mastering hedging in English.Skelton（1988）and Bloor&Bloor（1991）find that unqualified writing is more typical of EFL students than native speakers.Hu，Brown&Brown（1982）and Allison（1995）find that Chinese L2 writers are more direct in tone by using more strong modals.Zhao（2009）and Zhao&Wei（2010）compare the stance devices produced by Chinese language learners and British students，indicating limitations in the traditional language teaching in China.

There is also a growing body of research that has sought to identify the rhetorical features of particular languages，which are then compared to those in English research articles.Surez-Tejerina（2005）analyses a collection of Spanish and English book reviews.Working with a Move Analysis，the writer finds considerable consistency in the two languages with respect to the occurrence and the expression of evaluation，but with some rather different attitudes from the underlying cultures as can be seen from the fact that English book reviewers are more likely to make negative remarks than Spanish reviewers.Moreno（1997）makes a contrastive analysis of the use of causal metatext in orienting readers in the interpretation of cause-effect intersentential relations（CEISRs）between RAs in English and in Spanish on business and economics written by native speakers of each language.The results show that both language groups seem to make CEISRs explicit with similar frequency and strategies.The only differences are in their tendencies towards verbal or nominal anaphoric and anaphoric-cum-cataphoric signals.Jiang and Tao（2007）compare hedges in discussion sections of English and Chinese and observe the differences in the types and frequencies of hedges.

Outside studies in the academic genres，an influential study is conducted by Crismore，Markkanen&Steffensen（1993）which investigates cultural and gender variations in the use of metadiscourse in the United States and Finland by asking whether U.S.and Finnish writers use the same amounts and types of metadiscoursal devices and whether gender makes any difference.They find that students in both countries use much more interpersonal than textual metadiscourse with Finnish males using the most and U.S.males the least.Milne（2003）explores the role of metadiscourse in the editorials of the Spanish El Pais and the British The Times.She finds not only the uniformity of interpersonal metadiscourse，but also interesting differences in the frequency distribution of textual vs.interpersonal metadiscourse.Tang（2006）finds that English writers use more hedges in preface-writing than Chinese writers do.

To sum up，most of these cross-linguistic studies compare the frequency distributions of certain interpersonal features（either lexical or grammatical），providing clues as to how culture may influence the expression of evaluation.The basic assumption for such studies is that reliable equivalence can generally be obtained between expressions in two languages.However，there seem to be two traps that we are likely to fall into.First，most words are polysemous and in most cases the comparison relies upon intuition or general translation competence.Their functions are often difficult to be pinpointed，which consequently hinders cross-linguistic analysis.Second，even in the case of seemingly reliable equivalence（such as we and我们），the function of the items can only be determined by looking at the context.