4 minute read
English
Getting to know... Miss Veronese
1. What is your favourite book?
ARG. You can’t ask an English teacher that! It’s not fair. So I’m going to cheat, and give you a top five. For childhood comfort, I would choose The Little Prince and Moominsummer Madness. From an academic perspective, I would have to go for the spectacularly bizarre Beware the Cat, a sixteenth-century Reformation beast fable featuring horseriding, homicidal cats and an unfortunate incident involving a pair of walnut shoes… My two other very special books are The Passion by Jeanette Winterson and The Waves by Virginia Woolf.
2. Which 3 fictional characters would you want to be stuck in a lift with?
Shakespeare’s Feste would probably make some mischievous entertainment, whilst also singing a few songs which would expose the true nature of everyone in the lift. So that would pass the time. Then we’d need a good raconteur – maybe Chaucer’s Wife of Bath? She’d be a laugh. Although the situation would be a good opportunity to scrutinize an enigmatic character – so maybe one of literature’s shadier folk, like Cornelia Funke’s Dustfinger or Milton’s Satan.
3. Who would play you in the biopic of your life?
Helen Bonham Carter - I adore her insanity. She’s a curly girl, and my wardrobe would spectacularly improve.
4. You’re new at Notting Hill, what are your first impressions so far?
From the word go, I’ve been struck by the genuine enthusiasm and independence of the NHEHS girls. Arriving at interview, I was told the Sixth Form had decided to start their own feminist book group. I’ve encountered students inspired by Dylan Thomas to write villanelles, and Year 7s disappointed to be reading only four pages worth of excerpts from Emmeline Pankhurst’s ‘Freedom or Death’ rather than the entire speech. Any one of those moments would make my entire year, but at NHEHS they happen every day, and that’s pretty extraordinary.
Machine Translation with Marcus Tomalin
In September, Marcus Tomalin from Downing College, Cambridge, came in to give a presentation on machine translation. He discussed the fascinating overlap of mathematics and computer science with linguistics and literature. The crossovers between these supposedly directly opposed subjects was certainly refreshing to explore.
After a brief rundown of the history behind machine translation, Tomalin explained ‘Neural Machine Translation’. This is the framework that Google Translate and other translation services use, and it works by using a complex systems patterns and probability, mapping words to numbers in order to assemble correct syntax structure and choosing the vocabulary based on data it had previously seen.
There are different ways in which the reliability and skill of a network is measured, including interrogating the accuracy and fluency. Tomalin acknowledged the great strides that machine translation has made, but also the many ways it is still lacking, particularly in accuracy. For instance, Google Translate often uses pivoting (instead of translating directly from French to Icelandic, which it might not have much data on, it’ll translate from French to English and then finally Icelandic) which effectively means that an error is twice as likely. Then, there are many problems with the data that is being analysed. With language being the strange social art that it is, one word may carry many different connotations depending on context. This means different networks may be suited to translating different things depending on the data they have seen, so one network is better at translating business contracts while another is better at translating tweets. And then, because translation services learn and mimic the data they are exposed to, some of the prejudices society holds end up being reflected back at us. Tomalin gave the example of ‘nurse’ being automatically gendered as female in a translation, because in most of the data the program is trained on, nurse is given as female.
Of course, as well as the troubling insight into the various stereotypes and biases all cultures have and how they are manifested in language, there is also a huge chance of comic mishaps. Machine translation really struggles with words with multiple meanings as well as metaphors and other literary devices. This means that humour and poetry often translate poorly. Though this can show the many interesting differences between languages and cultures globally (how many languages have the phrase ‘it’s raining cats and dogs’, and in how many languages is that ’mumbojumbo’), it is often just amusing. Tomalin introduced the work of Marzia Grillo, who took poems from Emily Dickinson and translated them instantly into Italian, and then published the results. Dickinson’s ‘One with the banner gay’ (‘gay’ here meaning joyful) became ‘Uno con la bandiera gay’ (‘gay’ here meaning homosexual, as opposed to ‘felice’ which means happy).
Machine translation seems to be still in its infancy, but it improves with every day and is gradually learning to accurately translate sophisticated texts. Even without machines, perfect translation and loss of meaning were always issues of debate and controversy, and so whether a perfect translation is possible really depends on your stance on those issues. In the end, Marcus Tomalin managed to shed some light on a complex and relevant topic, helped bridge gaps between subjects, and keep a discussion about machines quite human.
