Saturday, October 13, 2012

Delighted Beauty: An Interview with Tom Cheesman

Dr Tom Cheesman at Swansea University runs a fantastic website, Delighted Beauty, on the multiple translations of great writings, and the differences between the translations, and what that can tell us about the work in question.

 You can try out the site/tool here and you can read about it in this article.

 How to use a 2-minute introduction:

 • Sign in top right of home page, as guest (instructions provided). Select Corpora > Othello corpus > Open.

 • On the right of the ‘Base text’ pane, the buttons lead to various visualisations. (Another is accessed by clicking any name in the corpus.)

 • First choose ‘E & V’ (Eddy and Viv): and for a good example of what’s on offer, locate Brabantio’s ‘Here is the man, this Moor’, click on it, and then try the ‘Sort by’ and ‘Order’ options above.

Dr Cheesman agreed to answer a few questions for Brave New Words.

Q: Why did you decide to start this project?

A: A few years ago I was working on a very interesting and controversial trans-adaptation of Othello, by Feridun Zaimoglu and G√ľnter Senkel. They said in an interview that they’d looked at “more than a dozen” other German translations. I’d no idea there were so many. Thinking of comparing their translation with earlier and contemporary ones, I started collecting books and scripts. Soon I had about 40: far too many! Studying multiple translations throws up problems of presentation, scale and navigation. They’re all acts of interpretation both of the translated work and of their own cultural contexts; they differ in countless details, all of which can be significant and interesting; it’s difficult to present a study of such a mass of texts without demanding far too much of your reader. You can study how a tiny sample gets differently translated; I did that, with one couplet, and I also created a website for crowd-sourcing the same couplet in any language. But really I wanted to find a way to explore whole texts and whatever aspects we want to. In 2010 my colleague David Berry convinced me of the potential of computational methods. Amazing work such as Ben Fry’s Traces made me realise that web-based digital tools, especially data visualization techniques, can help anyone explore multiple translations of anything, any way they want. Instead of being piled on shelves, digitised translations are available for anyone to look into (subject to ©). Text data visualisations (textviz) can work with full text contents (like the ‘parallel view’ on our site), or with representations derived from texts (using non-words, like our ‘alignment maps’, or words, as in things like Wordles). So textviz can bypass language barriers to some extent: I can find things out about Chinese translations compared to Russian ones, without knowing either language. In our ‘Eddy and Viv’ view, we use machine translation, which cries out for a user correction and discussion option, in a Wiki or whatever: still, it’s useful. -- ‘Eddy and Viv’ was a key conceptual breakthrough: at some point I worked out that formulae from information retrieval and stylometry could be adapted to put a mathematical value on the ‘distinctiveness’ of a translation in relation to others (a lot of retranslations are often similar, but some are more dissimilar than others); and that, if you first break a translated text into segments (sentences, play-speeches, or whatever), and then calculate distinctiveness values for all the translations of each one, then you can derive a value for the amount of variation among translations which a segment provokes. Where in a text do translators disagree more and less? Which speeches, which character parts? That’s the idea behind the tool we call ‘Eddy and Viv’.

Q: Was it a challenge to find funding? What about to find collaborators/contributors?

A: The idea might take a few minutes to explain but is basically simple and powerful, and funders like that. I’ve been very lucky with collaborators here in Swansea. David Berry put me in touch with a Computer Scientist, Bob Laramee, a specialist in Data Visualization. He’d not worked with texts before but was intrigued by the idea, and he and his PhD Zhao Geng have done some great work – as yet offline. Then we came across some fascinating text visualisation work by Stephan Thiel, doing just the kind of things we were interested in: ‘Understanding Shakespeare’. Obviously here was a designer we could work with. And Stephan was up for it. Our university’s Research Institute for Arts and Humanities funded a pilot project – an RA, Alison Ehrmann, copied and scanned my collection of German Othellos. We had a really good break when I requested a free trial of OCR software from ABBYY, and their UK sales manager Colin Miller got in touch personally and offered to help. He gave us free use of their Recognition Server, with Gothic fonts recognition and a suite of historical and modern dictionaries. And finally, along came the AHRC with their ‘Digital Transformations’ funding opportunity, just as my head of department, Andy Rothwell, recruited Kevin Flanagan to do a PhD here on Translation Memory. Kevin just happens to have all the skills and attitude needed to build a corpus management database and segment/align tools from scratch in order to make a ‘Translation Array’ actually work, feeding data to cool interfaces designed by Stephan at Studio Nand. I can’t code for toffee. It could all have gone horribly wrong.

Q: What has been one of your most interesting findings/realisations?

A: Lots! At the level of ‘Tell me about German Othellos’, my one-couplet sample study showed that in the 1950s and 60s, translations briefly became more distinctive or perhaps ‘daring’ than ever, then more conservative than ever, and then around 1990 they became really quite wild. The Array confirms that overall pattern but with some adjustments. Now the translator Hermann Motschach, in the 1990s, turns out to be right at the top of the scale, or to put it another way, his version is wildly wilful in almost every line. There is no public information at all about Hermann Motschach, except that he has translated almost all of Shakespeare’s plays and his scripts are used quite a lot by German theatres. Needs studying!

I found out a lot of interesting things when I researched the short descriptions I wrote for each translation. The stories behind them tell the 19th-, 20th- and 21st-century history of Germany and Austria. Check out Bab/Levy and Zeynek, for instance.

More broadly, I keep finding out about more and more examples of multiple translations, sometimes quite astonishing, all grist to the Arrays mill of the future. For example, Project Yao is a database of American fiction in Chinese, co-created by Joe Lockard, who tells me that some canonical novels such as Call of the Wild, published in 3 or 4 translations until about 1990, have since been published in up to 40 translations: that’s different translations, not re-editions. The sheer volume of versions of Shakespeare is crazy, too…

Q: What have you enjoyed the most about this project? And the least? A: The worst thing is that now that we’ve created this site with its array of Othellos, I have no time to use it for research, to explore the texts and find out how they vary and what that might correlate with. (Do translators translate Othello’s part more differently than other parts? is one question raised by what I’ve been able to do so far.) The priority now is to get more funding. When I’m not teaching, preparing teaching, or writing for you, BJ, I’m working on the next grant proposal.

There are many good things. Working with my collaborators. The delight of Jan Rybicki at seeing the platform: as soon as it was launched he started using it to work on his English-Polish novels corpus. The sense of exciting new horizons. What we have is a proof of concept prototype which just demonstrates the potential in a small way. This could become a big deal, not just for translation studies, comparative studies, language studies, cross-cultural studies, but beyond ‘studies’: for public understanding of translation and language, and for creative translation practice. If you’re doing a new translation/adaptation of a classic, and 300 versions are explorable online, in umpteen languages, with interpretative tools and a global social network around them, then of course you can ignore that. But you have some interesting new options too. 

Thank you so much, Dr Cheesman! And good luck with Delighted Beauty!

No comments: