What is this?
There are a number of digital editions of the Canterbury Tales online. Some of them take advantage of the web’s ability to provide glosses dynamically through frames or tooltips. This edition is conceptually a bit different. It reads much like a printed edition of the text with a marginal gloss. The dynamic feature is that the gloss can be changed according to the degree of difficulty of the language. Put another way, the reader can choose to gloss more or less based on how easy he or she finds the language. As the reader’s ability improves, he or she can reduce the amount of glossing.
Adjusting the gloss can provide a way for beginners to learn Middle English without losing confidence. It can also help them measure their progress. In addition, it can help reduce over-reliance on the gloss by allowing readers to test themselves by reducing or removing it.
Another issue with digital texts is that their glosses are generally only usable if the text is read online. This edition generates a printable version of the text with the amount of glossing selected by the reader.
How Does It Work?
The texts are stored as a matrix of Middle English word, Modern English transliteration, and gloss. The transliteration is simply an equivalent of the word in modern spelling if the word has survived in present-day English. It does not have to have the same meaning in Modern English as it does in Middle English. The original word and the transliteration are then compared by a simple Levenshtein distance metric. If the resulting measure similarity between the two falls below the threshold set by the reader, the gloss is displayed. If it does not, the gloss does not appear.
In other words, the measure of difficulty of the Middle English is the degree of difference between the spelling of Middle and Modern English equivalents. Where the Middle English has no equivalent in Modern English, the similarity is considered to be zero.
Naturally, comparing spellings is a somewhat simplistic method of establishing how easy it is for modern reader’s to understand Chaucer’s language. Once the initial form of the edition is complete, the plan is to devise more sophisticated measures. Ultimately, machine-learning techniques will be employed to establish this using crowdsourced or other user-generated data.
Development Notes
The dynamic gloss currently works on Google Chrome and recent versions of Internet Explorer. If you cannot see the form to increase the similarity threshold, please switch to one of these browsers.
All glosses are for the moment provisional. The following notes provide an indication of the stage of completion for each tale as of 4 December 2012:
- General Prologue: Initial draft complete. Misalignment of text and gloss corrected on 7 February 2013.
- The Knight’s Tale: Transcriptions and gloss produced by students in my Fall 2012 Chaucer class are provided without corrections. This may reflect both the similarity calculations and the reliability of the interpretive gloss. Please use with caution until corrections have been made.
- The Miller’s Tale: Initial draft complete.
The aim is to continue producing tales with student help. A full list of student participants is provided on the credits page.