Hugh Craig

Shakespeare, computers, and the mystery of authorship

1. Introduction Hugh Craig and Arthur F. Kinney 2. Methods Hugh Craig and Arthur F. Kinney 3. The three parts of Henry VI Hugh Craig 4. Authoring Arden of Faversham Arthur F. Kinney 5. Edmond Ironside and the question of Shakespearean authorship Philip Palmer 6. The authorship of The Raigne of Edward the Third Timothy Irish Watt 7. The authorship of the Hand-D addition to The Book of Sir Thomas More Timothy Irish Watt 8. The 1602 additions to The Spanish Tragedy Hugh Craig 9. Transforming King Lear Arthur F. Kinney Conclusion Arthur F. Kinney Appendix A. Plays in the corpus Appendix B. A list of 200 function words Glossary.

Propositional Idea Density in women's written language over the lifespan: Computerized analysis

The informativeness of written language, as measured by Propositional Idea Density (PD), has been shown to be a sensitive predictive index of language decline with age and dementia in previous research. The present study investigated the influence of age and education on the written language of three large cohorts of women from the general community, born between 1973 and 1978, 1946-51 and 1921-26. Written texts were obtained from the Australian Longitudinal Study on Womens Health in which participants were invited to respond to an open-ended question about their health. The informativeness of written comments of 10 words or more (90% of the total number of comments) was analyzed using the Computerized Propositional Idea Density Rater 3 (CPIDR-3). Over 2.5 million words used in 37,705 written responses from 19,512 respondents were analyzed. Based on a linear mixed model approach to statistical analysis with adjustment for several factors including number of comments per respondent and number of words per comment, a small but statistically significant effect of age was identified for the older cohort with mean age 78 years. The mean PD per word for this cohort was lower than the younger and mid-aged cohorts with mean age 27 and 53 years respectively, with mean reduction in PD 95% confidence interval (CI) of .006 (.003, .008) and .009 (.008, .011) respectively. This suggests that PD for this population of women was relatively more stable over the adult lifespan than has been reported previously even in late old age. There was no statistically significant effect of education level. Computerized analyses were found to greatly facilitate the study of informativeness of this large corpus of written language. Directions for further research are discussed in relation to the need for extended investigation of the variability of the measure for potential application to the identification of acquired language pathologies.

Propositional Idea Density in aphasic discourse

Background: Measuring and describing the effects of aphasia on the informativeness of language is a complex process. Due to technological advances in the recent years, the processes involved in the measurement of language can be automated through the use of computerised analyses. In the present research, the Computerized Propositional Idea Density Rater (CPIDR 3.2) provides an automated method for calculating Propositional Idea Density (PD), a measure which has been shown to be sensitive to the effects of ageing and dementia. The measure of PD quantifies the proportion of words within a text that are semantically intrinsic to its overall meaning. Aims: This research investigated the extent to which PD measures were different in aphasic and non-aphasic discourse, and the extent to which PD correlated with the severity of aphasia and with the established measures of other aspects of informativeness. Given the previously reported high levels of agreement between the computerised analysis and human raters, it was hypothesised that there would be high levels of agreement between the computerised analysis and human judgements for aphasic (as well as non-aphasic) discourse. Methods & Procedures: Data from the Goals in Aphasia Project were analysed for the purposes of the present research. De-identified transcriptions of 50 interviews with individuals with aphasia and 49 interviews with their family members were stripped of all interviewer data, leaving only conversational contributions made by the participants. These formatted transcripts were analysed using two automated, computerised language analysis tools: CPIDR 3.2) and Systematic Analysis of Language Transcripts (SALT Version 8) for a range of other discourse measures. Outcomes & Results: Results showed a significant difference in PD (p < .001) between aphasic and non-aphasic discourse, and PD decreased significantly as aphasia increased in severity (p < .001). The concurrent validity of these findings was supported by the findings of relationships with established discourse measures. The total percent agreement between the computerised analysis and human judgments for aphasic discourse was 99.57% and for non-aphasic discourse was 99.74%. Conclusions: The findings indicated that PD has the potential to be used as a measure of discourse informativeness in aphasia and that further research into this approach to analysis is warranted.

Contrast and Change in the Idiolects of Ben Jonson Characters

The paper presents the results of a series of Principal Components Analyses of the frequencies of very common words in the dialogue of characters in plays by Ben Jonson. The first Principal Component in the data, the most important axis of differentiation, proves in each case to be a spectrum from elaborate, authoritative pronouncements to a dialogue style of reaction and interchange. Reference to other quantitative studies, literary and otherwise, suggests that a version of this axis may often be among the most important in stylistic difference generally. In Jonson it has a chronological aspect -- there is a shift over his career from one end to the other -- and there is often significant change within the idiolects of his characters as well. Successive segments of Volpone and Moscas parts (they are protagonist and antagonist of Volpone, perhaps Jonsons best-known comedy) change markedly along this axis, beginning far apart but coming by the end of the play to resemble each other very closely on this measure.

Language chunking, data sparseness, and the value of a long marker list: explorations with word n-grams and authorial attribution

The frequencies of individual words have been the mainstay of computer-assisted authorial attribution over the past three decades. The usefulness of this sort of data is attested in many benchmark trials and in numerous studies of particular authorship problems. It is sometimes argued, however, that since language as spoken or written falls into word sequences, on the ‘idiom principle’, and since language is characteristically produced in the brain in chunks, not in individual words, n-grams with n higher than 1 are superior to individual words as a source of authorship markers. In this article, we test the usefulness of word n-grams for authorship attribution by asking how many good-quality authorship markers are yielded by n-grams of various types, namely 1-grams, 2-grams, 3-grams, 4-grams, and 5-grams. We use two ways of formulating the n-grams, two corpora of texts, and two methods for finding and assessing markers. We find that when using methods based on regularly occurring markers, and drawing on all the available vocabulary, 1-grams perform best. With methods based on rare markers, and all the available vocabulary, strict 3-gram sequences perform best. If we restrict ourselves to a defined word-list of function-words to form n-grams, 2-grams offer a striking improvement on 1-grams. .................................................................................................................................................................................

Language and ageing – exploring propositional density in written language – stability over time

This study investigated the stability of propositional density (PD) in written texts, as this aspect of language shows promise as an indicator and as a predictor of language decline with ageing. This descriptive longitudinal study analysed written texts obtained from the Australian Longitudinal Study of Womens Health in which participants were invited to respond to an open-ended question about their health. The 635 texts used for this study were taken from 127 middle-aged women who responded to this question on each of the five surveys conducted at 3-year intervals over a 16-year period. The study made use of an automated PD rater (CPIDR-3) for the analysis. PD was found to be a stable measure over time when comparing the grouped data, but there was between- and within-subject variation over time. Further research is needed to explore the valid use of this measure in research into language and ageing.

Is the Author Really Dead? An Empirical Study of Authorship in English Renaissance Drama

For some time there has been debate in literary studies, and especially in the field of Shakespearean scholarship, about the importance of authorship in understanding and categorizing literary texts. In an analysis of affinities between 100 plays by various authors from the Shakespearean period, based on frequencies of very common words, authorship emerged as distinctly more important than genre or date in grouping plays. Cluster analysis showed further that, while authorial affinities are overwhelmingly dominant in the early stages of clustering, where only the closest pairings are considered, a small subset ofan authors plays typically remains apart from his other works as the analysis proceeds. The study indicates that in Shakespearean drama authorship is objectively detectable, and indeed very important, though it must also be acknowledged that these authors also regularly created texts which are not easily assimilable to the larger clusterings oftheir works.

Jonsonian Chronology and the Styles of A Tale of a Tub

Jonson’s career writing for the stage and the masquing hall was so long that the idea of innovation and change in his style was commonplace even in his lifetime. In the 1630s commentators referred to a decline in his powers since the glory days of Volpone and The Alchemist, anticipating Dryden’s idea that these recent works were Jonson’s ‘dotages’ (Jonson being then around sixty). Indeed, the existence of an extensively revised version of Every Man In His Humour is evidence that by 1615 or 1616 Jonson himself was conscious that his notions about dialogue, character, plot and even setting had altered in the 15 or so years since the play had been first composed.

George Chapman, John Davies of Hereford, William Shakespeare, and A Lover's Complaint

In this essay the author reassesses affinities between the language of the poem and the characteristic linguistic habits of Chapman, Davies, and Shakespeare. He uses a single large corpus of poems and plays by these three writers and a number of their contemporaries, as well as some straightforwardly quantitative measures: a list of words peculiar to an author, a list of words never appearing in his or her work, a list of adjacent word pairs that appear regularly in an author’s work, and a list of pairs appearing rarely in it. The methods are tested by treating a series of samples whose authorship is known as if they were anonymous. The results for Chapman and Davies are stark. Their work can be readily distinguished, and A Lover’s Complaint is quite unlike these authors’ writings. The answer for Shakespeare is less clear cut—“cross-validation” of the results shows that the methods are less successful in providing unambiguous separations - but the poem has some marked affinities with his general practice.

Who wrote 'A Visit to the Western Goldfields'?: Using computers to analyse language in historical research

Determining the authorship of unattributed writings can be a major issue for scholars. As this article demonstrates, computational stylistics provides a valuable methodology in helping to answer the question, ‘Who wrote it?’ Gold occupied much space in the newspapers of colonial Australia in the 1850s-70s. It kept many reporters very busy. Few, however, are known by name. An exception is Charles de Boos, a prolific reporter for the Melbourne Argus and especially the Sydney Morning Herald. Whilst it is possible to identify much of his work, questions arise over the authorship of other columns, such as the series ‘A Visit to the Western Goldfields’. Stylistic analysis has confirmed that this series is not the work of de Boos, but that of another writer who remains anonymous. No methodology answers every question, but this example illustrates the potential of computational stylistics to be an important aid in many areas of historical research. This article has been peer-reviewed.