High-frequency verbs in the ICLE Corpus
Håkan Ringbom · 1998
The ICLE corpusThe International Corpus of Learner English (ICLE) provides excellent material for studying advanced learner English (see Granger 1994).A comparison of these argumentative essays from 15 different countries with each other and with a corresponding native speaker corpus can shed light, for example, on the relationship between general characteristics of learner language and Ll-transfer.Two key terms in lexical corpus analysis are 'overuse' and 'underuse'.These terms are not ideal, since they presuppose a norm from which one subcorpus deviates.Neither professional writing, as exhibited for example in the belles-lettres sections of the LOB or Brown Corpus, nor the British and American undergraduate essays in the ICLE native speaker corpus, provides an ideal norm of comparison with advanced learner writing.In fact, the whole idea of a norm of NS competence is based on the existence of 'an ideal speaker-hearer' in Chomskyan terms.For essays on topics varying as much as they do in the ICLE corpus, the idea of a norm of argumentative writing is difficult to envisage.The ICLE NS essays probably provide the least unsuitable material -student writing after all seems a better alternative than professional writing -but American and British student essays naturally exhibit a number of linguistic and argumentative idiosyncracies and shortcomings.What can always be done, however, is make comparisons between NS and learner essays, without using the native essays as the yardstick against which learner writing should be measured.If we, then, for want of a better alternative, accept the ICLE native speaker corpus as a basis for comparison, we still have to be wary about the terms 'overuse' and 'underuse' whenever we find that one particular learner corpus shows clear frequency differences from the native speaker corpus.It is clearly not enough to consider a learner corpus only in relation to native speakers writing similar types of essays: it should also be compared with other learner corpora.Only if frequencies in one subcorpus deviate in the same direction from both the NS corpus and several other learner corpora can the terms 'overuse' and Håkan Ringbom -9789004653658 Downloaded from Brill.