The Pear Tree Project: a geographico-statistical and linguistic analysis

The original analysis of the figures relating to the Pear Tree project (see Salway 2007) involving detailed breakdowns of the results recorded at universities in Catalonia, Ireland, Italy, Poland and Belgium, have already provided interesting, if rather limited, data relating to how film texts are perceived by subjects coning from different nations/languages/cultures. These figures have now been supplemented by results recorded in Germany, Greece, South Africa, Britain, America, Spain and France. These additions, which have been integrated with the original figures, will give greater scope for meaningful examination using the tools of multivariate statistical analysis. This will consist of discriminating sets of variables and identifying a distribution gradient with the aim of recognising sets or cases with similar characteristics. The codified variables will be collected in a single matrix to establish percentage frequencies. Through clustering and factorial analysis, it is hoped to identify sets of variables that will assist in the process of highlighting and understanding any differences in the various descriptions of the Pear Tree text. Furthermore the figures will be subjected to analysis following the canons of corpus linguistics. The Pear Tree texts form a small corpus in themselves and this will be studied in terms of word frequency, clustering, concordance and colligation. The two types of analysis are in many ways complementary and it is hoped that the results obtained from both will provide assistance in understanding perception variance across cultures, with a view to fulfilling one of the main aims of the project, namely to glean information necessary to create optimum audiodescriptions of audiovisual texts for blind communities everywhere.