Os quantificadores a few e few: questões de interlíngua e prosódia semântica em corpus de aprendizes

Abstract

The investigation described here was carried out under the light of what has been done by researchers (GRANGER; TRIBBLE, 1998) who are interested in learner English writing. The main goal was to verify and describe the upper-intermediate Brazilian learners interlanguage, specifically concerning the use of the quantifying expressions a few and few. Data from Br- ICLE, the Brazilian subcorpus of the International Corpus of Learner English (ICLE) was analyzed so as to enable description both on the grammatical level and on the level of semantic prosody. Similarly, data from the Louvain Corpus of Native English Essays (LOCNESS), which was built to be the native equivalent for ICLE, was analyzed and used for comparisons. It was also necessary to resort to the academic section of COCA as another reference corpus in order to describe the way native speakers use the quantifiers being studied. The data from both Br-ICLE and LOCNESS were treated qualitatively, through manual analyses of the expanded context of concordance lines whose node word was few. The following sorts of collocations were identified and compared: the collocations that appeared adequate, that is, compatible with native speakers use; and the ones that appeared inadequate, that is, showing deviations that were further systematized and discussed. The type of inadequacy that received special treatment was the one related to the surfacing of semantic prosody. In the native speakers corpus, the collocation very few was often used to express negative semantic prosody; however, this was not attested in the Br-ICLE, which leads to the understanding that such collocation was underused by the learners in similar contexts. Quantitatively, the results of this research were described in percentages, and the diagnosed underuse was validated through statistical testing. In order to gauge the extent to which the mother tongue influences the expression of semantic prosody in the target language, another corpus, whose size is on a par with Br-ICLE and LOCNESS, was compiled: the Co-RAP, which stands for Corpus of Academic Essays in Portuguese (Brazilian Portuguese). Also, the online platform of Corpus do Português was used for further necessary searches. The conclusion drawn was that, in their writing, Brazilians do not seem to widely use muito poucos(as), which is the correspondent collocation in Portuguese to very few, and this may influence their lexical choices when writing in English. Lastly, following the tenets of Data- Driven Learning (JOHNS, 1994), a classroom activity that encourages the recognition of lexico-grammatical patterns was suggested, and promising pathways to future research were mentioned as well.