Corpus-based evaluation of pedagogical materials: If-conditionals in ELT coursebooks and the BNC

摘要

This paper aims to contribute to the growing body of what may be termed pedagogy-driven corpus-based research; that is, research which is situated at the intersection of language description, pedagogical grammar, and pedagogical materials evaluation (e.g. Kennedy, 1992; Owen, 1993; Hunston & Francis, 1998; Harwood, 2005; Römer, 2004, 2005). The paper is based on an ongoing study of conditional sentences in the British National Corpus and examines the representation of the typology used in a number of English language teaching (ELT) coursebooks, with reference to a random sample of 1,000 if-sentences from the written sub-corpus of the BNC. The paper reports on the frequency of the ELT types in the sample and outlines the cases that the ELT typology does not cover. It also focuses on the use of modality and modal expressions, as well as what ELT treats as ‘special cases’, such as ‘if + Past tense’ with past time reference, and (semi-)fixed expressions (e.g. if any, if not, if so). Overall, current ELT coursebooks essentially use the typology in Logic (real, counterfactual and hypothetical), which they term first, second and third types respectively, with the addition of two more types, zero and mixed. In addition, they adopt a restricted (and restricting), if not naïve, approach to modality, in that only central modals feature in the definitions and examples for each type. Normally, ELT materials present specific combinations of verb forms and modals in the two clauses, as well as the time reference and attitude to likelihood that each combination, rather than each clause, expresses. That is, they instruct learners what combinations of tense-aspect marking and modal auxiliaries to use in the protasis and apodosis, as well as what attitude to possibility and time reference each resulting combination expresses. They also outline ‘special cases’ which the ELT typology presumably does not cater for. A number of studies have reported that the ELT typology fails to account for a large number of attested if-conditionals, and provides learners with a narrow and inaccurate view of if-conditionals (Hwang, 1979; Maule, 1988; Fulcher, 1991; Wang, 1991; Ferguson, 2001). The study findings corroborate these conclusions, while providing more detailed quantitative information. However, this study differs from those cited above in two respects. It adopts a modular approach to analysis; that is, the if-clause and main clause of the conditionals in the sample were annotated manually for, on the one hand, the tense-aspect marking of the main verb and the modal expressions, and, on the other, the modality marking and time reference. Also, the study examines the ELT typology on its own terms, and distinguishes four levels of inclusion, determined on the basis of the information given in a sample of ten coursebooks for advanced learners. Previous pilot examinations of ten ELT coursebooks for upper-intermediate and advanced learners published between 1998-2002 (Gabrielatos, 2003a,b, 2004) showed that even if no distinctions were made on the basis of the semantic/pragmatic relation holding between the two clauses, and even if the typologies given in all ten coursebooks were conflated into an inclusive one, this typology would only account some 44% of the cases found in the sample. The introduction of the mixed type does not seem to be an educationally sound decision, as this type showed a mere frequency of 1.5% in the sample, and it would become obsolete if a more data-based approach to a pedagogical typology were adopted. The distinction between the traditional first type and the newly introduced zero type, too, proves to be unnecessary and confusing, as the difference between the two has nothing to do with the presence/absence or nature of modalization in the two clauses. Rather the distinction is one of time reference; specifically, the distinction is between timelessness (zero) and future reference (first). It seems educationally sound, then, to conflate the two in one category, corresponding to what Quirk et al. (1985) term open conditionals. One of the most significant limitations of the ELT typology, arguably inherited from the typology in Logic which it is based on, is that it distinguishes types only in terms of modality marking and time orientation. More specifically, it ignores the category of conditionals termed indirect conditionals (Quirk et al., 1985), speech act conditionals (Sweetser, 1990), or pragmatic conditionals (Athanasiadou & Dirven, 1997). This paper will examine current editions of the same coursebooks (when applicable) or current coursebooks of the same type and level as those used in the pilot studies. It will report on any adaptations in their treatment of conditionals - thus also examining to what extent ELT coursebooks seem to become informed of corpus findings. It will also provide a more detailed analysis, which also takes account of the semantic/pragmatic relation between the two clauses. The paper will conclude that the present treatment of if-conditionals in the ELT coursebooks examined is characterised by the following basic shortcomings: a. It provides learners with an incomplete, and in some cases distorted, picture of if-conditionals, in terms, on the one hand, of their morphosyntactic patterns and modality marking, and, on the other, the link between morphosyntax and the semantic/pragmatic relation between the clauses. b. It tends to overwhelm learners with long lists of ‘special cases’ or 'exceptions' (when rules of some description are given), or by offering a piecemeal account of possible combinations without showing how they may fit a coherent framework. c. It potentially limits the learners' language production by restricting their repertoire to a small number of pre-fabricated combinations of protasis and apodosis, which, in addition, may not be among the most frequent ones.