2007 Fiscal Year Final Research Report Summary
Multivatiate Approaches to Dickens' Style
| Project/Area Number |
17520324
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
English linguistics
|
|---|
| Research Institution | Osaka University |
|---|
Principal Investigator |
TABATA Tomoji Osaka University, Graduate School of Language and Culture, Associate Professor (10249873)
|
|---|
| Project Period (FY) |
2005 – 2007
|
| Keywords | stylistics / corpora / statistics / multivariate analysis / stylometry / Dickens / English prose / linguistic variation |
|---|
| Research Abstract |
Examination of methodological issues : Various types of word frequency profiles have been compiled from the electronic corpus of texts which were collected and digitized in the present and the past research projects. The types of data include high-, mid-, and low-frequency word variables. Several word-classes are also used as variables, such as-ly adverbs and superlatives. By applying correspondence analysis, the present research has revealed how sharply texts are differentiated between authors as well as how texts are clustered according to chronology within authorial sets. The multivariate approach demonstrated in this research allows examination of the complex interrelationships among texts, interrelationships among linguistic variables, and association between the texts and linguistic variables graphically in a multidimensional space. It computes the row coordinates (text scores) and column coordinates (word scores) in a way that permutes the original data matrix so that the correl
… More
ation between the word variables and text profiles are maximized. In a permuted data matrix, linguistics variables with a similar pattern of distribution make the closest neighbours, and so do texts of similar profile. When the text/variable scores are projected in multi-dimensional charts, relative distance between variable entries indicates affinity, similarity, association, and other similar relationships between them.
Multivariate approaches to the study of Dickens' style : The most powerful solution obtained from this methodology clearly differentiates between the Dickens and control sets. Of special interest is that the authorial differentiations have emerged in almost any experiments conducted, regardless of the variables employed. What is also striking is that a variation from Dickens' early works to later works was identified in most of multivariate analyses. Of further interest is that, in the Dickensian set, serial fiction texts stand in marked contrast to texts in other genres, such as sketches and history. The series of results illustrate how the authorial difference, text genre, and chronology are reflected in the frequency pattern of linguistic variables in the texts written by Dickens and the control authors. The results also suggest the effectiveness of the multivariate approach to stylistic studies of texts. Less
|
Research Products
(33 results)
