Names in novels: An experiment in computational stylistics.

Proper names in literary texts have different functions. The most important one in real life, identification, is only one of these. Some others are to make the fiction more ‘real’ or to present ideas about a character by using a name with certain meanings or associations to manipulate the reader’s expectations. A description of the functions of a certain name in a certain text becomes relevant when the researcher can point out how it compares to the functions of other names and names in other texts. The article describes how research into names in literary texts needs a quantitative approach to reach a higher level of relevancy. To get a first impression of what may be normal in literary texts, a corpus of twenty-two Dutch and twenty-two English novels and ten translations into the other language in both sets were gathered. The occurrences of all names in these novels have been tagged for those data categories that seemed useful for the literary stylistic research planned. Some first results of the statistics are presented and the use of the approach is illustrated by means of an analysis of the use of geographical names in the Dutch novel Boven is het stil by Gerbrand Bakker and its English translation by David Colmer, The Twin. In the evaluation of the results, special attention is paid to the status of currently available digital tools for named entity recognition and classification, followed by a wish-list for the tools that this kind of research really needs. [ABSTRACT FROM AUTHOR]